Archive Magazine CD 1995/Archive Magazine CD 1995.iso/text/hints/volume_6

Text File | 1995-06-25 | 401.2 KB | 8,902 lines

Hints and Tips
6.1
•   Alt characters − On upgrading from RISC-OS 2 to RISC-OS 3.1, I thought
I had lost my ‘×’ character. I eventually found that it had moved from
<alt-,> to <shift-alt-,>. (Even on RISC-OS 3.0 on the A5000 it is still
<alt-,>.) So I investigated all the possible ALTernatives(!). Here are
all the ALT characters I found. They are laid out in QWERTY order but
where no special character appeared, the key is not included in the
list.
6.1
key   ALT   shift ALT
`   ¬   °
1   ¹   ¡
2   ²
3   ³
4   ¼
5   ½
6   ¾
9      ±
r   ¶   ®
y      ¥
o   ø   Ø
p   þ   Þ
\      ¦
a   æ   Æ
s   ß   §
d   ð   Ð
f      ª
z   «
x   »
c   ¢   ©
m   µ   º
,      ×
.      ÷
/      ¿
6.1
I think that is all there is, but if anyone finds any more, let us know.
Ed.
6.1
•   Annual service kit − In answer to Paul’s comment about hard drives in
Archive 5.12, p.8, I’d like to sing the praises of the Alarm application
(on one of the support disks or in ROM on RISC-OS 3). (You will see the
link if you read on! Ed)
6.1
I, too, would have forgotten completely about the annual replacement of
filter and batteries, but for this application (which, thanks to my Boot
file, always resides on the iconbar): When my machine was a year old,
Alarm reminded me of this fact in no uncertain terms, (‘urgent’ alarm,
repeating every hour), so there was very little chance of overlooking
this necessary bit of maintenance. Jochen Konietzko, Köln, Germany.
6.1
•   Apple Laser Writer II NT − Here is some information for anyone wanting
to make a cable link from an Archimedes to an Apple Laser Writer II NT.
The connections are as follows...
6.1
   Archimedes   LaserWriter II NT
   Pin         Pin
   2   (RXD)   −−−−   3   (TXD)
   3   (TXD)   −−−−   2   (RXD)
   5   (0V)   −−−−   7   (0V)
   6   (DSR)   −−−−   20   (DTR)
The Archimedes pins DCD (1), DTR (4) and CTS (8) are all linked
together.
6.1
When using a RISC-OS printer driver you should choose the Serial
connection with 9600 baud rate, 8 data bits, no parity and 1 stop bit.
XOn/XOff protocol should also be used.
6.1
•   DDE Pascal − Some would-be users of DDE !Pascal (on Careware 17) will
have discovered that it won’t function in Computer Concepts’ Compressed
Filing System (CFS), where it generates the error ‘Special field not
allowed in filesystem prefix to OS_CLI’. This means that the command
line interpreter, which is invoked to run the Pascal tool, objects to
some pathname which begins with the prefix ‘CFS#’. The offending
pathname turns out to be the <Pascal$Dir>.Pascal, as assigned into
Alias$Pascal by the !Run file with the apparent function of enabling the
compiler to be located in the application directory. The remedy is to
comment out the line which assigns Alias$Pascal and move the compiler
(i.e. <Pascal$Dir>.Pascal) to your library. This brings the configura
tion into line with that of the C compiler and other DDE tools and
causes ‘*Pascal’ commands to be serviced by default from the current
library without any explicit reference to a filesystem.  Mike Kinghan,
Oxford
6.1
•   DDE Pascal (Careware 17) − The release of the public-domain Norcroft
Pascal compiler for the Archimedes is potentially extremely useful,
since, although ISO-Pascal is very good, it does not interface well with
the operating system. More seriously, programs compiled with ISO-Pascal
do not interface well either.
6.1
The Norcroft system is easy to use and install and does everything that
is claimed of it. Unfortunately, users should be aware of some problems,
especially since Acorn obviously does not wish to maintain the product.
6.1
I have some extra severe tests for Pascal, which I have used to test the
compiler. This shows that the compiler has some errors in the evaluation
of set expressions: the following program failed (printing True rather
than False):
6.1
{4 5 6 4 0 3    3000 }
{the parameters from the full test which failed. }
program test (output);
const
   ten = 10;
var
   i: 1..ten;
begin
i := 11; (compilation error here?)
i := ten + 1; (compilation error here?)
writeln(‘Error not detected‘) ;
end.
6.1
The first error is caught at compile-time by almost all Pascal com
pilers, including ISO-Pascal. The second error is caught by almost all
compilers at run-time, and a few at compile-time.
6.1
When using the quit option in the Pascal menu, the system still allows
command line usage of Norcroft Pascal, although the icon is removed.
6.1
The final significant problem is that programs written in ISO-Pascal do
not necessarily compile without the ISO option being set. This is
annoying, since many programs will wish to use some extensions. I don’t
know of any other Pascal compiler for which the extension option means
less (in some cases).  Brian Wichmann, NPL.
6.1
•   Directory size − Regarding the issue on directory sizes, etc in the
hints & tips section, there are a few other points...
6.1
RISC-OS 2 has some problems with certain combinations of disk size and
allocation unit, so beware. The problem usually causes the machine to
hang when entering a directory. I have only experienced this using large
allocation units on small disks.
6.1
The fact that a new dir appears to take 38Kb isn’t the whole story, as I
believe that there is a facility for small files within a directory to
use some of the ‘wasted’ space. However, if you tend to have lots of
directories with few files, space does tend to get wasted.
6.1
The best solution as far as both disk, space and speed is concerned is
to use D format, which uses the ‘correct’ amount of space for direc
tories and hardly anything for the map. (Remember that there are two
copies of the map in E format.) D format disks are also MUCH easier to
recover files from if the root directory or map gets corrupted and files
can never be fragmented (which slows down access). The only disadvantage
is the occasional need to compact but I find this a small price to pay.
I would only recommend E format for non-technical users.
6.1
The same issues will apply to all FireCore based filling systems such as
IDE, etc.  Mike Harrison
6.1
•   Impression/Squirrel − CC’s !Importer (also in their Business Supple
ment) is supplied by Digital Services with Squirrel. Here is some
experience that might help other people.
6.1
Using Squirrel, I created a database with 22 fields. I then tried to
mail merge the resultant report CSV file using !Importer. I got an error
message saying ‘Subscript out of range (Code 6150).’ After some hours of
fiddling, I eventually found that if I reduced the number of fields in
the CSV file to 16, !Importer worked OK. Nick Horn, Thetford.
6.1
•   Second internal hard drive in an A410/1 − I fitted a 40Mb ST506 hard
drive to my A410, thinking it would be more than adequate. However, with
the acquisition of the PC Emulator and the formation of a 10Mb DOS
partition, I was rapidly running out of room. I subsequently purchased a
second hand 20Mb ST506 drive, which I decided to fit internally rather
than buy a p.s.u., case and fan. I thought of making some mounting
brackets but, having seen an advert for a SPACE kit from PRES which
included brackets and all the leads, I changed my mind and went for
that. Price was £38.72 inc VAT. The work involved was to:
6.1
1)   Remove top two podule blanking plates.
6.1
2)   Fit metal brackets to drive using screws supplied.
6.1
3)   Remove the 34 way data cable and replace it with the one supplied,
connecting it to the existing drive and mother board (SK11), leaving
spare connector to connect to the second drive later.
6.1
4)   Connect the supplied 20 way control cable to the spare connector
(SK8) on the mother board, leaving the other end to be connected to the
second drive later.
6.1
5)   Remove the power cable from the psu and existing drive and replace
it with the supplied cable, leaving the spare connector for the second
drive.
6.1
6)   Lead all three cables to the back of the computer and through the
space left by removing the blanking plates and connect them to the
second drive.
6.1
7)   Attach the blanking plates to the metal brackets on the second
drive using nuts & bolts supplied.
6.1
8)   Ensure second drive is set to ‘1’ by changing the appropriate DIP
switch or DS-Jumper.
6.1
9)   Insert the drive through the back of computer until the front of
the brackets rests on backplane connectors (psu side) and the blanking
plates line up with the back of the computer. Screw in blanking plate
screws.
6.1
10)   Once computer is up and running it will have to be configured for
a second hard drive.
6.1
The work was very easy to carry out and the instructions supplied were
very good. The drive then has to be formatted using Acorn’s HFORM
program. The program had to be modified to do the ‘slow’ formatting by
altering the following lines:
6.1
7530 REMmed out
7540 REMmed out
7550 REM removed
7560 REM remove
6.1
Once this was done, the drive formatted with no problems and, using the
PC Emulator’s configure program, I created a 19 Mb DOS partition.
6.1
I would recommend this solution of gaining extra drive capacity if you
have access to a cheap drive and do not mind losing the space of two
half width podules.  Chris Bass, Grimsby
6.1
•   Mah Jong from CIS − If you have recently purchased the Mah Jong game
from Cambridge International Software, you may wish to check for the
presence of a ‘benign’ virus. This can be done by double-clicking on the
!MahJong game while holding the <shift> key down. Choose the ‘Display >
Full Info’ option on the filer window. If the MemAlloc module is larger
that 2608 bytes then your copy is infected and you should contact CIS
directly at Unit 2a, Essex Road, London  N1 3QP or telephone
071−226−3340.
6.1
•   Measuring free space − I would like to say a big thank you to all the
people who wrote in response to my plea for help in finding a way of
checking, from within a program, whether there is enough space for a
file before it is saved.
6.1
The most ingenious suggestion was to use the following commands:
6.1
*FREE {ram:ccc }
X=OPENUP(“RAM:CCC”)
REPEAT B=BGET#X: UNTIL B=ASC“=”
REPEAT B=BGET#X
 IF B>47 AND B<58 B$=B$+CHR$B
UNTIL B=ASC“B”
*SHUT
6.1
to put the free space in the variable B$.
6.1
Some people pointed me to ‘ADFS_FreeSpace’ (also RamFS_FreeSpace,
SCSIFS_FreeSpace, etc.) which can be used with syntax such as:
6.1
SYS “ADFS_FreeSpace”,Drive$ TO Free%,Largest%.
6.1
The same information was given on page 7 of the Archive 5.12.
6.1
Although all these answers have been very helpful, they have also taught
me a salutary lesson, namely that you have to be very precise when
describing your problem: because, although they have answered the
question (and told me things that I did not know) they haven’t actually
told me what I needed to know!
6.1
What I actually want is a filing system independent way of finding out
whether a file with any given path name can be saved. For instance, if
the user of my program indicates (by dragging an icon to a filer window
or typing into a save box) that he wants to save a file with the path
name ‘XXXFS::$....Filename’ and if my program knows how long the file
will be, I want my program to be able to find out whether the file will
fit in the specified place, without knowing in advance whether ‘XXXFS’
stands for ADFS or CFS or SCSIFS or RamFS or whatever. The program also
has to work if a path is not specified and the file is simply to be
saved to the CSD.
6.1
A couple of correspondents have pointed out that there is an SWI called
‘FileCore_FreeSpace’ (on page 1021 of the PRM). However, in order to use
this you have to know the ‘private word’ of the filing system in
question and, so far as I can see, this SWI can only therefore be called
by a filing system module, which is not much use to a mere mortal
programmer!
6.1
I still, therefore, believe that the only generalised way of doing what
I want is to use syntax like the following :
6.1
X%=OPENOUT(Filename$)
SYS “XOS_Args”,6,X%,Length% TO ,,Allocated%
IF Allocated% >= Length% THEN ... go ahead and save it.
6.1
This is hardly elegant but it seems to work.
6.1
I understand that RISC-OS 3 will include the generalised SWI that I
want. However, this will not be much use, of course, in a program that
is designed to work in RISC-OS 2 as well!  Hugh Eagle, Horsham
6.1
•   Formatting numbers (continued) − Over the last two months, I have been
following the correspondence on Basic’s number formatting ability (PRINT
and STR$) with some interest. Has the increased size and complexity of
RISC-OS caused the oversight of some of the slightly esoteric but very
powerful features of BBC Basic? One such feature is the @% variable in
Basic. The @% variable controls the format of the PRINT and STR$
functions and much of last month’s discussions could have been settled
by altering the @% variable so that STR$ was formatted in the same way
as PRINT.
6.1
@% is fully discussed in the manuals that accompany the computer under
the PRINT command but I will try to give some useful pointers here. As
with all Basic integers, the @% variable is four bytes long. When Byte 4
(the left-most byte) is zero Basic ignores @% when using the STR$
function (this is the default and the source of last month’s problems)
otherwise STR$ is formatted exactly the same as PRINT. Byte 3 allows the
choice of three different formatting types: 0 − general (the default), 1
− exponent and 2 − fixed. Byte 2 specifies the number of digits printed
and Byte 1 gives the print field width when you use comma separators in
PRINT.
6.1
Our problem arises from two areas: the way computers represent numbers
and the value of @%. At this point, I think it is necessary to take an
aside into the nature of real number representation in computers so, if
you are only interested in how to circumvent Basic’s printing problems,
skip to the end of this article.
6.1
Common advice to novice programmers is not to use real variables unless
you really have to. Speed and storage reasons are normally given for
simplicity but a far more important reason is that real numbers are not
exact in the same way as integers and they must be treated differently.
A classic example is the problem in mathematics of dealing with two
representations of, say, the number one: which could be 1.000... or
0.999.... As far as we are concerned both are the same but if we ask a
computer whether or not they are equal, we find not. This is due to no
fault of the computer or the programmer but the acceptance that
computers cannot store the infinite expressions for real numbers but
must approximate to a number of decimal places and so springs up the
subject of numerical analysis.
6.1
How then do we equate real numbers? Well, we must choose a value for a
zero! That is a number where any number smaller than it we consider to
be ‘as good as zero’. An example for Basic is 10−7, any smaller and the
errors in calculations become significant. So when we want to check
equality of two numbers a and b we use the expression (ABS(a­b) < 1E-7).
If this is true then the two numbers are ‘as good as equal’. (ABS is
used since we are only interested in the magnitude of the difference
between the numbers.)
6.1
The second important point with numbers stored in computers is that they
are stored in binary. Yes, of course, you all knew that but fewer people
are aware how computers store fractions. Much like in decimal when we
have a decimal point, computers use a binary point where the digits to
the right of the point have place values 1/2, 1/4, 1/8, etc. In decimal
there is a problem of representing fractions like 1/3 or 1/6 because the
prime factors of 3 or 6 are not divisible by the factors of 10 (2, 5) so
you get infinite decimal expansions for them. In binary, the problem is
greater because there is only one factor of two (unlike two factors of
ten) so even more numbers have infinite representations (this is why
some people advocate using base 12 instead of base 10 for day to day
life − it makes mental arithmetic so much easier). This means that
simple decimal numbers like 0.6 have an infinite expansion in binary.
6.1
What has all this got to do with printing numbers in Basic? Well, an
appreciation of what is going on when you store numbers will help you
avoid the pitfalls that last month’s contributors highlighted. Now that
we are aware that innocuous looking numbers like 2.1 can have infinite
expansions when stored in binary, it becomes clear that 2.1 × 10 may not
equal 21 but, perhaps, 20.99999999!
6.1
Is there anything that we can do about this? Yes, indeed there is. We
can tell Basic to round off the number to nine decimal places instead of
attempting to print 10 which extends the precision beyond the accuracy
(in general Basic is accurate to nine decimal places unless you are
using the floating point version where I believe 19 is the limit but not
having a RISC-OS 3...).
6.1
Now we come back to @%. At default, @% specifies 9 decimal places but it
only affects PRINT and not STR$ which defaults to ten. To alter this,
all we need to do is set the byte 4 flag in @% to one. Try PRINT
STR$(0.6) and then set @% = &0100090A and try again (by the way, &90A is
the default value for @% whereas STR$ uses &A0A if byte 4 is unset).
6.1
If we want to print money then we can set @% to two decimal places by
&0002020A (remember to set byte 4 to one if you want to use STR$). This
has the added bonus of allowing you to print the number rounded off but
still maintain the extra places for further calculations.
6.1
I hope you can now see the power and simplicity of using @%. My only
warning is that if you wish to use several formats in your program, then
either save the previous @% or declare it as a local variable.  Emmet
Spier, Guildford.
6.1
This subject seems to have opened up quite a debate. Emmet has agreed to
edit any comments regarding rounding errors for us and so if your wish
to add to the debate you should write to him directly. Emmet Spier, 6
Parklands Place, Guildford, Surrey  GU1 2PS.
6.1
•   PC keyboards for A4 − A word of warning − not all PC type keyboards
work on the A4 (or all PCs for that matter!), so take the A4 to the shop
and try before you buy! Reasonable ones can be had for about £40 at some
computer fairs. Note that non PS/2 type keyboards (with the 5 pin DIN
plug) can be used with an adaptor or by changing the plug.  Mike
Harrison
6.1
•   RISC-OS 3 DOS partitions − Because DOS filenames (including exten
sions) are in general longer than the maximum length of a valid ADFS
filename (10), it is not always possible to copy or move DOS files from
hard or floppy discs directly to the ADFS.
6.1
The !MultiFS utility as supplied with the PC Emulator solves this by
allowing you to truncate the DOS filename while the translation process
takes place. Another very useful option (which I often use when copying
files from C programs) is to force !MultiFS to handle DOS extensions as
(hierarchical) directories.
6.1
However, when the filetype of a DOS partition on your ADFS hard disc is
set to ‘DOSdisc’, RISC-OS 3 will display a filer window for the
partition when you try to open the folder, while !MultiFS just ignores
it (i.e. no :C drive on the iconbar.) This is all very nice but RISC-OS
3 doesn’t offer the nice name translation options mentioned above. To
enable !MultiFS to display your DOS partitions, just set their filetype
to ‘data’ before you start !MultiFS. Remember to use the command line
and not the filer menu because the latter doesn’t allow you to change
the the filetype of DOSdiscs. I inserted 3 lines for each DOS partition
in the !MultiFS !Run file to swap between ‘DOSdisc’ and ‘data’ type
partitions when !MultiFS is started:
6.1
   *Filer_CloseDir    <PCe$Drive_X>
6.1
*SetType    <PCe$Drive_X> Data
6.1
directly after setting the system variable <PCe$Drive_X>, and
6.1
   *SetType    <PCe$Drive_X> DOSDisc
6.1
at the end of the !Run file (although the comment tells us not to do
so).
6.1
Unfortunately, there are some bugs in !MultiFS: When you copy to a
MultiFS filer window, you must ensure name translation is not set to
hierarchical or you will end up with wildcards (‘?’) in filenames. When
this happens you must use DOS itself to delete the file because RISC-OS
fails to wipe those objects. Another bug appears if you shut down your
system after using !MultiFS. This results in an error report and a task
manager crash.  Paul Groot, Holland.
6.1
•   Saloon Cars Deluxe − If you are having problems running Saloon Cars
Deluxe, it may be worth knowing that you must have the Font Manager
installed for it to run.
6.1
•   Shareware 44 (Fortran Friends) update − There is a small bug in the
!Fortran77. DesktopF77 program. Line 1400 should read:
6.1
1400 IF INSTR(b$,“rror”) THEN E%+=1
6.1
This allows errors in the code generator part of the compiler to be
detected. Kate Crennell, Oxon.
6.1
•   Zelanites cheat − The following program provides a cheat for
Zelanites. It stops the number of lives decreasing below 9.  Steve
Edwards, Stourbridge.
 10 REM >:0.$.hints.zelcheat
 20 REM Archive hints & tips section
 30 REM (c) S.EDWARDS 1992
 40 REM infinite lives for Zelanites
 50 REM ****************************
 60 REM * When you have run the program the game should *
 70 REM * then be run as normal. This time when you die  *
 80 REM * you will have not gone down one life but up   *
 90 REM * to NINE lives, which stay with you for ever.  *
100 REM *                          *
110 REM ****************************
120 :
130 DIM store% &400
140 MODE 0
150 PROCDisc_op
160 PRINT‘********FINISHED********’
170 PRINT
180 PRINT‘****NOW GET BLASTING****’
190 END
200 :
210 DEF PROCDisc_op
220 PRINT ‘Put !ZELANITE disc into drive 0’
230 PRINT ‘   and press a key.’
240 A=GET
250 PRINT ‘Loading  sector information’
260 SYS ‘ADFS_DiscOp’,0,1,&50400, store%,&400
270 store%?&67=&6A:REM Original code =&2A
280 INPUT‘Are you are ready to make the change ?’;A$
290 IF LEFT$(A$,1)=‘N’ OR LEFT$(A$,1)=‘n’ THEN END
300 IF LEFT$(A$,1)=‘Y’ OR LEFT$(A$,1)=‘y’ THEN
310   PRINT‘Saving cheat back to disk’
320   SYS ‘ADFS_DiscOp’,0,2,&50400, store%,&400
330 ENDIF
340 ENDPROC

Hints and Tips
6.2
•   3D icons for RISC-OS 3.1 − Here are some tips for those of you who
upgraded from RISC-OS 3.00 to RISC-OS 3.10 (and for anyone else who is
interested in a fancy desktop).
6.2
If you have used the fancy 3D-like window borders on your A5000, you
will probably have noticed that, with RISC-OS 3.1, the borders will not
install properly. This is because RISC-OS 3.0 programs used the
‘IconSprites’ command to install the wimp’s tool sprites (i.e. borders),
while in RISC-OS 3.1 there is a new command to install user defined
window tools:
6.2
*ToolSprites <spritefile>
6.2
This is, I think, because the border sprites are now held in a separate
sprite area for faster redrawing. So, the only thing you have to do is
to change the command ‘IconSprites’ into ‘ToolSprites’ at the appropri
ate places in your files. Note that you can also re-install the default
window tools by entering ‘ToolSprites’ on the command line.
6.2
The small utility I use to install enhanced icons (a modified version of
!SetIcons), shows another minor modification of RISC-OS 3.1. It is now
assumed that some icons have the same dimension. For example, some
border elements are painted from two sets of sprites − one that is
normally used and one that is used when you select this particular
border element. If corresponding sprites within these sets are different
in size, you will see that some parts of the elements are not painted
when you select them. To overcome this problem, you can adjust the sizes
with !Paint.
6.2
For those programmers who liked the Impression-style dialogue boxes,
there exists a module called ‘Interface’. This module takes care of most
of the code to generate fancy icon borders. RISC-OS 3.0 has some in-
built extras to achieve the same thing without the need to insert extra
lines of code into your program. (The programmer only has to define some
validation strings which contain the information to generate the desired
effect.) However, because the validation string command syntax was
essentially the same as the syntax used by the Interface module, there
were sometimes problems when using the Interface module with RISC-OS
3.0.
6.2
After I installed the new RISC-OS 3.1 ROMs, I was surprised to see that
my own (RISC-OS 3.0!) programs didn’t have fancy borders any more. After
some experimenting with the validation strings, I discovered that the
command syntax is now different from the syntax used by the Interface
module. To be more specific, RISC-OS 3.1 makes use of ‘r<int><colour>’
(where <int> is a number between 0 and 8 and <colour> is a WIMP colour
number), instead of ‘b<int>, ...’. This modification probably has been
made to overcome the incompatibility problems with the Interface module.
The simplicity with which you can use this prompts a question: Why don’t
the ROM based applications like Draw use this facility? I guess this is
just another ‘Acorn-mystery’.  Paul Groot, Holland.
6.2
• Editing paper sizes in RISC-OS 3 − If attempts are made to edit the
paper sizes contained within the standard RISC-OS 3 printer driver
files, it will be found that any changes made to the Text Margins
section will revert to 0 when the window is closed, even if Save
Settings is selected. This is due to a bug, but can be rectified by
editing the paper files directly.
6.2
Editing any of the paper sizes results in a new file called PaperRW
being generated in the Printers application directory. This file can be
modified using Edit. Make sure that the Printers application is not
running (i.e. Choose Quit from the iconbar menu if it is). Find the
!Printers application (normally in the App1 directory) and open it using
Shift-double-click. Open the file called PaperRW by Shift-double-
clicking on it. Scroll down the file until you find the name of the
paper setting which you defined or modified. It will look something like
’pn: Fanfold A4’. At the bottom of the set of numbers relevant to this
paper setting are five values (called tb, tt, tl, tr, th) which
correspond to the text bottom, top, left and right margins and the
height (in lines) respectively. Alter the values from 0 to your required
numbers and save the file. When you re-run the Printers application,
your paper sizes will be rectified.  Simon Moy, N.C.S.
6.2
•   How NOT to treat your SCSI hard disk − A few weeks ago I found a new
way to get rid of all those files that clogged my external hard disk(!)
6.2
Every verify I did showed up one questionable sector, so I decided to
map it out as defective. The *Defect command applies only to ADFS, so I
browsed through the SCSI disk’s manual and, finally, found what I was
looking for (or so I thought).
6.2
Morley supply their hard disks with a formatter application which,
amongst other things, can add defects to a list. There is a warning that
a low level format would result in total loss of data but, later in the
description, an option was mentioned to decide whether or not to do the
low level format while running the application.
6.2
This led me to believe that clicking on ‘No’ would keep my data intact −
alas, I was wrong!
6.2
The Morley hotline informed me that I had in fact done a high level
format which wipes out the information about the addresses for each
file. As the passage in the manual could be misunderstood, they very
kindly offered to restore my data for me free of charge (just p&p).
(Great service!!)
6.2
However, it was not to be: I forgot to inform them that I use Computer
Concepts’ Compression on most of my files and this, I was told later,
encodes the files in such a way that retrieval becomes virtually
impossible.
6.2
I learned two things from my self-made disaster:
6.2
1) Even the excellent Compression is not proof against all kinds of
fools.
6.2
2) 800 Kb floppy disks are just not good enough as a backup medium! I
shudder at the thought that I will have to feed 112 floppies into my
machine, once the hard disk returns from England.
6.2
As a consequence, for future backups I have ordered a 20Mb floptical
drive from Morley, to be sent as soon as it is able to read and write
Acorn floppy formats.  Jochen Konietzko, Köln, Germany.
6.2
•   Humour in RISC-OS 3.1! − You will be pleased to hear that humour still
exists within Acorn. If you load ‘Resources:$.Resources.
Switcher.Templates’ into the template editor, FormEd, you get an
interesting message.  Knut Folmo, Norway
6.2
•   Loading multiple Impression documents − If you want to load a number
of Impression files at once, just select them (with rubber band dragging
on RISC-OS 3, if you like) and drag them down onto the Impression icon
on the iconbar.
6.2
• Memory size from Basic − In Basic, there is no specific way to find
out how much physical memory is fitted to the host machine. However, it
can be calculated using a SWI call as in this example:
6.2
10 SYS “OS_ReadMemMapInfo” TO PageSize%, NumOfPages%
6.2
20 TotalMemory% = PageSize% * NumOfPages%
6.2
30   REM This value is in bytes
6.2
40 Total% = TotalMemory% / 2^20
6.2
50   REM This value is in Mbytes
6.2
Simon Moy, N.C.S.
6.2
•   More RISC-OS 3.1 Alt Characters The current A5000 Welcome Guide
mentions some additional Alt-key characters. These alt-keys don’t
actually produce a character themselves, they just set up an accent, so
that the following character is changed into an accented character. This
only happens if the accent is suitable for the following character.
6.2
For example to type an ‘e’ with an acute accent (é): Whilst holding down
the Alt key, press and release <[>. Release <alt> and the press and
release <E>. Thus:-
6.2
Alt-[  E     é
Alt-]  E     è
Alt-;  E     ë
Alt-’  E     ê
Alt-,  N   ñ
Alt-.  A     å
Alt-/  C     ç
6.2
The only other Alt-key character missing from the previous list is Alt-
hyphen which gives character 173, which, in the Acorn fonts, looks
exactly like a normal hyphen, but is often different in fonts from other
suppliers.  Simon Moy, N.C.S.
6.2
• MS-DOS 5, PC Emulators and the Archimedes keyboard − (This excerpt
from a reply to an Archive reader’s letter may prove useful to others.)
The cursor keys do not work correctly with the PC emulators due to a
bug. They will not work properly with either QBasic or DOS Edit. They
can be used to move around the text but, to select text, you must use
the cursor and editing keys which are mapped on the numeric keypad (as
on a normal PC AT-style keyboard). Therefore, when inside the editors,
make sure that NumLock is off and use <shift> and the following keys (on
the numeric keypad):
6.2
6.3
6.2
AMOUSE.COM is fully compatible with MS-DOS 5. However, as with all PCs,
you cannot display the mouse pointer at the DOS prompt. Also, not all PC
programs support the mouse. However, if the AMOUSE command is present in
your AUTOEXEC.BAT file and the program you want to use accepts a mouse
(e.g. QBasic, EDIT and DOSSHELL), you must start the appropriate
application first and then select Connect Mouse (by pressing <menu> over
the PC emulator window) if you are running the emulator from a window in
the RISC-OS desktop. If you are using the emulator under Single Task
mode, the mouse should automatically connect itself. (If the mouse does
not connect itself automatically when running the emulator under Single
Task mode, then this is a further bug in the PC emulator.) Typing AMOUSE
at the DOS prompt or selecting Connect Mouse from the window’s menu will
not display the DOS mouse pointer, as DOS itself does not support a
mouse. If you do this, the RISC-OS pointer will disappear (because
control has passed to DOS) and you must press <menu> on the mouse to
relinquish control from the emulator back to RISC-OS.  Simon Moy, N.C.S.
6.2
• Pound sign problems − The pound sign has always been a difficult
character to print and display on the screen. To make the pound sign
appear in the PC Emulator, press and hold <alt> and type 156 on the
numeric keypad. When you release <alt>, the pound sign will appear.
6.2
If you are printing a file as text (i.e. from Edit or PipeDream for
instance) then the pound sign will print as a hash (#). You can reverse
this by using a hash symbol in your text in place of the pound sign,
although it is important that your printer is set up with a UK character
set − see the DIP switch settings for your specific printer. If you are
printing a file produced using bitmaps (e.g. Impression) then the pound
sign will print correctly anyway. Basically, any program which defaults
to printing on the screen using the system font will print characters as
plain text − there is a section on plain and fancy text in the RISC-OS
User Guide.
6.2
By the way, does anybody know why you can’t access the pound sign
(either from the keyboard or <alt-156>) in a Task Window under RISC-OS
3.0?  Simon Moy, N.C.S.
6.2
• RISC-OS 3.1: A310 − There is a problem when fitting RISC-OS 3.1 to
A310 machines fitted with the Arcram 434 ARM3 + 4Mb upgrade from Watford
Electronics. It causes the machine to hang on power-up. Replacement PAL
chips and more information are available from Watford Electronics on
0923−37774 (Fax: 33642).  Simon Moy, N.C.S.
6.2
• RISC-OS 3.1: SCSI drives − There may be problems accessing the SCSI
drive using the HCCS Hard Card interface with RISC-OS 3.1. Contact HCCS
for details.  Simon Moy, N.C.S.
6.2
• Rotated fonts and sprites using Laser Direct − With the advent of
RISC-OS 3.1, there are problems with printing rotated text and graphics
using the currently existing RISC-OS 2 printer drivers such as that
supplied with Laser Direct. To overcome the problem, transform the text
or sprite to any angle in Draw, make sure that the item is selected and
choose the Convert to Path option from the Select submenu. This will
allow rotated text or sprites to be printed correctly using the RISC-OS
2 printer driver supplied with Laser Direct. This does not mean that it
will work with all RISC-OS 2 printer drivers − you will have to
experiment! Steve Costin, Ramsgate.
6.2
•   TV interference − In response to Chris Bass’ enquiry in Archive 6.1
p36, here is a general procedure which covers almost all cases:
6.2
1. Does the computer interfere with your own TV, when it is as close/
closer to your computer than your neighbour’s TV is? (On either side of
a party wall they could be less than 1 m apart!)
6.2
If YES, go to 2. If NO, go to 3.
6.2
2. Does the interference disappear when the computer is 3 m or more from
the TV?
6.2
If YES, agree with your neighbour to move your computer and his TV so
that they are more than 3 m apart. If moving them apart is inconvenient,
you could try sticking cooking foil on to the party wall (under the wall
covering, eventually). You need a minimum of about 2 m by 2 m, and there
is usually no point in earthing it, but you could try the effect.
6.2
In some types of building, your computer might be too close to your
neighbour’s TV aerial, in which case moving them apart is the best bet.
6.2
If NO, the computer may be faulty. Many dealers may not be able to cope
with this type of fault, so you, or your dealer, should contact Acorn
with a FULL description of the problem.
6.2
3. If NO to question 1, your neighbour’s TV has either an inadequate
aerial or inadequate ‘immunity’. (This is a technical term describing
‘resistance to electromagnetic disturbances’.) An inadequate aerial
usually also gives ‘fuzzy’ pictures and/or ghosts (outlines of images
displaced sideways). The aerial itself may be OK; the fault may be in
the downlead. They tend to fill up with water and deteriorate after a
time. A new aerial (and downlead) is quite inexpensive, even a large
one. If the aerial is not the problem, the manufacturer of the receiver
should be contacted, either directly or through the TV dealer or rental
company, regarding the probable lack of immunity of the TV. Manufac
turers (of TVs and computers) are quite concerned about this at present,
because of an EC Directive on the subject.
6.2
4. If all else fails, the DTI Radio Investigation Service will look at
the problem and give advice but there is a fee of about £21 involved.
The Service has much information and experience of such problems and
their cures, including private information about know immunity problems.
For Lincolnshire, the address is (or was) PO Box 2500, Nottingham NG8
3SS. The head office is at Waterloo Bridge House, Waterloo Road, London
SE1 8UA. A book, ‘How to improve Television and Radio Reception’ is/was
available from the Service, and gives a good deal of information on
interference problems (some of it technical).  John Woodgate, Rayleigh,
Essex.  A



Hints, Tips and Letters
6.2
A combined selection this month, since the hints and tips stem from
letters I have received recently.
6.2
• Key Window / Impression problem:  I am grateful to Jack Evans of
Bristol for pointing out this problem to me, particularly since I am the
author of the program in question. Key Window (available on Shareware 47
– see review in Archive 6.1 p31) is a program which displays keystrips
in a window, and the strips update themselves as you move the caret
between tasks.
6.2
The problem is that if you have Key Window running at the same time as
you are editing a document in Impression which is set to auto-save, the
auto-save doesn’t work. You can still save manually of course, but this
is something which is easy to forget about. The solution is easily
effected, if not very satisfactory. You have to de-select ‘Sense caret’
in Key Window’s Options menu. Unfortunately, this will prevent Key
Window from changing keystrips as you move between applications. If you
use Impression’s auto-save a lot, and want to disable Key Window’s
interference permanently, you can edit the !Run file to disable the
‘sense caret’ option via a system variable. Full instructions for doing
this are provided in the file.
6.2
Users may be interested to know that I have just started work on version
2 of Key Window, and am re-writing it from scratch to make it a much
better program with enhanced facilities. I intend to approach Computer
Concepts about the problem, to see if a satisfactory solution can be
found. Other programs with auto-save facilities, such as Ovation and
Vector, do not appear to suffer from the problem.
6.2
• Protecting Impression files:  Charles Martin from the Isle of Wight
has been having a few problems with some Impression ‘template’ files –
empty documents with standard formats. He wants to make them Delete- and
Write-protected via the filer’s Access menu but has found that setting
the files (which are, of course, in the form of application directories)
to ‘locked’ does not protect their contents. Similarly, opening the
directories and choosing ‘Select all’ does not have the desired effect.
6.2
This situation actually applies to all applications and directories, not
just Impression files. However, under RISC-OS 3, the Access menu is
completely redesigned and works much better, so the problem only really
affects RISC-OS 2 users. The difficulty arises because of the distinc
tion that the filer makes between directories (be they normal
directories or application directories) and actual files. This is one of
those ‘features’ of RISC-OS 2 which needs a bit of care.
6.2
Under RISC-OS 3, if you select a directory or application directory
(such as an Impression file) and set it to ‘Protected’, the filer will
protect all the files within it, no matter how many sub-directories
there are. Under RISC-OS 2, however, it will only set the status of the
directory itself. You therefore need to open the directory, select all
the files within it and set their attributes via the Access menu.
However, there is a further complication: if the selection of files
contains any directories, the process will be upset − the directories
need to be deselected first.
6.2
Under RISC-OS 2, to protect an Impression file, you would need to open
it up, choose ‘Select all’ from the filer menu, deselect any directories
in the selection by clicking on them with <adjust>, and finally set all
the files’ attributes via the Access menu. Then repeat the process for
each of the directories which you had deselected (‘MasterChap’,
‘Chapter1’, etc). There is simply no point in protecting the directories
themselves, since it is impossible to delete a directory which contains
locked files, and locking a directory (under RISC-OS 2) does not affect
the access status of the files within it in any case.
6.2
All in all, it’s probably easiest to upgrade to RISC-OS 3!
6.2
• Applause for Ovation:  A highly enthusiastic letter about Ovation has
been received from Tony Greenfield of Birmingham. He has the following
to say:
6.2
“[Ovation] does just what I had hoped for and more. The manual is
brilliant! It is a work book on DTP, quite good enough as an introduc
tion to the subject to beginners and as an encouraging tutorial for more
experienced users. I found it much more useful than First Impression. I
have learnt much about Impression II and much about how a manual should
be presented. I wonder if the reason that there are not many hints and
tips on using Ovation [in Archive] is because the user manual is so
clear? (It may also have something to do with the relative numbers of
people using the two packages. Ed.) There are many features which I like
about the program. In fact, there is not very much missing – I wonder
if, for most people, Impression is not really required − after all, it’s
more expensive than Ovation. I’m sure that, for schools who have yet to
purchase a DTP package, Ovation would be a must better buy than
Impression, for many reasons − perhaps most of all for the ease of use,
and I guess it would be easier to learn than Impression.”
6.2
Tony goes on to list various features which he likes in Ovation, notably
its wide range of effects such as word underline, small caps and
tracking, and its system of basing font styles on one family name, so
that changing the base family will also alter the italic and bold fonts.
6.2
I would certainly agree with Tony’s comments about the manual, which is
indeed a very helpful document, although I think Impression is the
easier of the two to learn, as Ovation has a quite unwieldy menu
structure. Perhaps the quality of the manual really is the reason why we
haven’t received more tips on using the program from readers. Anyway,
things are set to change from this month in Archive, with the advent of
Maurice Edmundson’s dedicated Ovation column. (We also have a new DTP
editor, Ray Dawson, whose first column will appear next month.) For
further comments about Ovation, see my Ovation/Impression comparison
article which follows on page 33.  A

Hints and Tips
6.3
•   A3010 joystick ports − several A3010s have been returned to Acorn with
complaints of faulty joystick ports. They are not faulty. The A3010 is
the first Archimedes to have a built-in digital (i.e. Atari-style)
joystick port and because of this, many of the games already available
do not cater for digital joysticks. This is a fairly major problem and
hopefully someone will soon produce some sort of patch (either hardware
or software) which allows digital joysticks to be used in place of the
more common proportional joysticks (such as the Voltmace Delta). Until
then, two sources of information about digital joystick interfaces are
The Leading Edge and Beebug.
6.3
•   A4 Portable Guide − on page 31 of this guide, the pin-out diagram for
the AKF18 monitor is slightly incorrect. Pins 13 and 14 (horizontal and
vertical sync) have been reversed.
6.3
•   Artworks & Impression dongles − There appears to be a clash between
these two dongles when using an A5000 and a JP-150: printouts appear
with spurious information across them. Computer Concepts say that the
problem only occurs between A5000s and JP-150 printers and add that they
are looking into the problem.
6.3
•   Banner-making − I have just seen, in Archive 6.1, a review of a
program called Banner.
6.3
If you own a printer which takes fanfold paper, there is an alternative
way for making banners. (By the way, fanfold paper does not need
sprocket holes if you are careful about the line-up and select friction
feed, so normal paper rolls can be used. For example, the Colourjet 132
users can print banners in this way.)
6.3
The print margins on all the Acorn printer drivers are user-definable
and as long as you want to print out banners shorter in length than A0,
extending the appropriate print margin and setting Draw to A0 landscape
will normally allow satisfactory banners to be produced. (Changing the
print margins is described in the manuals; change only the height of the
page to whatever you wish and leave the width alone.) If print margins
are shown, keeping to these will ensure that none of the picture is
lost.
6.3
The additional flexibility of Draw, especially in conjunction with the
drawfile utilities (!FontFX, !DrawBender, etc) would, in my opinion,
offset most of the ease of use of Banner, except perhaps in primary
schools.
6.3
Note that the printer should not be set to skip perforations, or
problems will result, but I have had no difficulty in printing banners
on an Epson LX and even a Colourjet 132 in this manner.
6.3
For those requiring longer banners, !Draw½ and, I believe, !Draw+
(Careware 13) allow an option to set the work area to the print margins,
which will then enable them to use the longer pages which the printer
driver can handle but, again, for a banner you should set the page to be
much taller than normal in the printer driver and edit the display in
landscape mode. I have used this to produce banners several metres long.
6.3
I am sure Banner is a very useful package and I have not attempted to
use it but I would be unwilling to spend £15 when the same effect seems
to be available from the software supplied with the machine and
Careware.   Andrew Garrard, Cambridge.
6.3
•   Banner-making again − Recently I needed to print up a large banner for
a meeting, and wondered how I should go about it. Of course, I could
always have bought ‘Banner’ for £15 but, for a once off job, I thought
this just a touch extravagant. This set me thinking... Impression can
define master pages as big as you like... hmm... This is what I did, and
it works!
6.3
1.   Open a new document in Impression
6.3
2.   Create a new Master Page (Custom Job) − Width: 1000 and Depth: 210
6.3
3.   Return to the document and set the Master Page in Alter Chapter
6.3
4.   Set the Font Size to 500
6.3
5.   Type in “Hello World!” or whatever
6.3
6.   Set the printer driver page size to − Height: 1000 and Width: 210
6.3
7.   Set the Impression Print Dialogue to print sideways, and fire away
6.3
Warning − this is not ‘printer ribbon friendly’... in fact, it is
positively ‘printer ribbon hostile’!  Philip Lardner, Co. Wicklow,
Ireland.
6.3
•   CC BJ10ex Turbo Driver − It seems that the only method of using the
Turbo Driver with the PC Emulator is to load !BJ10e and quit it again
before loading the emulator.  Chris McIntosh, Blackheath.
6.3
•   Programming the Serial Port (Vertical Twist) graphics board − On the
monthly program disc is a series of programs which, in themselves, do
nothing much. However, they show how the 65000 colour modes can be
accessed from Basic and C.
6.3
They mainly show how to draw lines, circles and fill shapes. Most do not
need any interaction. Fill requires five mouse selections to create to
outline − the sixth should be inside the box.
6.3
Lines − just move the mouse and press <select> to change colour.
6.3
15_120ASM converts mode 15/21 sprites to mode 120. Copy the sprite image
to the floppy’s root directory before running. It’s not very fast. Jim
Bailey, Sunderland.
6.3
•   Squirrel update − Digital Services have upgraded Squirrel from version
1.07. Currently, Squirrel is being supplied with the original V1.07 disc
and manual and a V1.20 upgrade disc. A new manual is due out soon and
Digital Services will send all existing Squirrel owners this new copy
providing they have registered.
6.3
•   VIDC enhancer & RISC-OS 3.10 − There can be problems using very early
Atomwide VIDC enhancer hardware with RISC-OS 3.10. It appears that if
the enhancer is software switched, RISC-OS 3.10 will automatically
switch it on which prevents the computer from starting up correctly. To
solve the problem, you have two choices: (a) at the command line, type
*FX 162,&BC,&80 and switch your machine off, then on again, to check
that the command has worked (full details are on page 14 in the RISC OS
3.10 release note) or (b) return the enhancer hardware to Atomwide with
£5 and they will modify the hardware so that it works correctly.
6.3
•   Vision Digitiser. As I concluded in my review of the monochrome
version of this product (Archive 5.12.42), the package is amazing value
for money. However, the supplied software is a little basic. Further
more, Robert Golightly at HCCS tells me that the software module
(promised in the !ReadMe file) to provide * commands and SWIs for users
to write their own programs has a priority below that of the 512-line
upgrade, and is not likely to appear in the near future.
6.3
All is not lost, for the software – at least the monochrome version – is
written using Basic’s in-line assembler, and so can be modified by the
careful user.
6.3
For example, as it stands, the !Vision application always demands 450Kb
of memory. For those with 1M machines, or those who like to multi-task
it with other memory-demanding applications, it would be useful to be
able to reduce this. Examination of the code reveals that a 320Kb screen
buffer is reserved, whatever screen mode you are using. 320Kb will be
needed for mode 21, but only 160Kb for modes 15 and 20, and 80Kb for
mode 12. If you haven’t a multi-sync monitor and want to save 160Kb of
memory when using !Vision, change the following lines of a copy of
!RUNIMAGE, having first checked that the first line of the !Run file
shows that you have Version 1.0, dated March 1992.
6.3
  130 DIM M% 210000
6.3
 8650 BNE mode_error
6.3
23260 EQUD 163840
6.3
23460 FN_work (163840)
6.3
Then amend line 2 of !Run to read:
6.3
wimpslot -min 288K -max 288K
6.3
I’m working on other possible improvements to the software, such as mode
67 compatibility, but it occurs to me that others might be doing the
same. If you are interested in participating in an informal user-group
for the Vision Digitiser, please send an SAE to Stuart Bell, 23 Ryecroft
Drive, Horsham, West Sussex RH12 2AW.  A
6.3
TechSoft
6.3
From 6.2 page 19
6.3
TechSoft
6.3
From 6.2 page 15
6.3
Risc Developments (2)
6.3
New Artwork
6.3
Using RISC-OS 3
6.3
Hugh Eagle
6.3
Many thanks to all those who have sent in material about RISC-OS 3.10.
Hugh has done a lot of work for us in collating all this information, so
read it carefully, check it out on your own systems and let us know if
you find solutions to some of the problems and/or see more problems.
This kind of information is very valuable indeed in terms of helping
other people who have to go through the same learning curve as you, with
the new operating system.
6.3
I think it would also be good to send in a few ‘I hadn’t realised that
you could... ...but it makes life much easier’comments.
6.3
By the way, as at 28/11/92, N.C.S. actually has the £50 upgrades in
stock although A5000 owners will still have to be patient, I’m afraid.
Ed.
6.3
(Throughout the column, references to RISC-OS 3 mean RISC-OS 3.10 unless
otherwise stated. HE)
6.3
After ‘The Long Wait’ and an acute dose of ‘Bent Pin Anxiety’, I had, to
cap it all, ‘The Hunt For The Missing Apps 1 Disc’. After all that, I
was quite surprised to find that the machine actually worked when I
turned it on!
6.3
Reconfiguring it so that it ran as well as it did before the upgrade,
was quite a fiddly process. The main hint I would give to anyone who
hasn’t yet done the upgrade is: watch your Fontmax settings!
6.3
•   In the Memory section of !Configure don’t be misled by the illustra
tion on page 94 of the manual. You can not safely leave the “Font cache
limit” setting at zero and expect the operating system to default to a
sensible limit as you can with most of the other memory settings. Unless
you set “Font cache” to a reasonably large amount, you must set a larger
“Font cache limit”. (On a 4Mb machine, I have them set to 128Kb and
512Kb respectively.)
6.3
•   If you have a multisync monitor, read the bit on page 119 (in the
section on screen modes!) when deciding what settings to use for anti-
aliasing and cached bitmaps in the Fonts section of !Configure.
6.3
•   Keith Lowe adds − Where the upgrade instructions tell you to hold down
the Delete key until the power-on self-test has been completed, they
mean don’t release the key until text has appeared on the screen!
Otherwise your old configuration will still be there and that can cause
no end of confusion!
6.3
•   Ed Harris encountered problems in installing the user fit ROM carrier
board from IFEL to his A310. IFEL were very helpful in providing advice
and when Keith was still unable to track down his mistake, they sorted
it out and returned it by return with only a nominal charge to cover
their costs. Many thanks to IFEL.
6.3
•   Mark Lambert reports problems trying to install RISC-OS 3 in an A310
fitted with a Watford 4Mb RAM upgrade (the second edition with a
vertical daughterboard holding the processor). Watford have told him
that a simple wire link is required on the RAM board but, as yet, they
have not told him the exact details. Does anyone happen to know the
details of the link required?
6.3
•   Steve Hutchinson writes: I was assured that it was quite simple to do
the necessary hardware upgrade to my old A440. In the event, I managed
it without hiccup but it was not quite so simple. I had to remove the
main circuit board from the case − which means unplugging loads of
things in my machine with all its upgrades. If you haven’t reasonable
electronic experience, think twice!
6.3
And is the upgrade worth it? For 1Mb or floppy disc only users, the
answer must surely be yes. It must be a godsend to have the standard
applications, plus fonts, plus modules, all in ROM. For hard disc owners
.... well, what do you think? Some readers have no doubts.
6.3
Ed Harris, for instance, writes: All in all, I have found RISC-OS 3.1 to
be a great improvement on an already excellent operating system (RISC-OS
2) and well worth the cost of the upgrade. However, I wonder whether
others share my disappointment that the improvements were not more
dramatic? Perhaps now is the time for us to begin to compile a Wish List
of things we would like to see in RISC-OS 4.
6.3
Why not send your thoughts either to Archive or to me at the address at
the end of this column?
6.3
Most programs work well in RISC-OS 3
6.3
Robert Macmillan of Colton Software has written to point out (further to
Gerald Fitton’s comments in the Pipeline Column of Archive 6.1) that not
only do all versions of Pipedream from 3.10 onwards work in RISC-OS 3,
they all work better with RISC-OS 3. This is because RISC-OS 3 has the
latest versions of system software built in and these perform better
than the old versions supplied with RISC-OS 2. In particular, the font
manager is much, much faster and the benefits of having the fonts built
in make it possible for floppy disc users to make good use of fonts
without inordinate disc swapping.
6.3
Robert suggests that Archive should publish a table showing which of the
most popular software packages work, and which don’t work or need
upgrading. Judging by the number of needless phone calls he has been
receiving, this would be of great benefit to users considering the
upgrade to RISC-OS 3.
6.3
I think the drawback with this idea is that although most people will
report that most programs work perfectly well (and the list could
potentially fill much of the magazine) some programs that do work well
for most people seem to cause problems for some others.
6.3
However, I do agree that it is important to emphasise that most programs
do work and so, to set the ball rolling, I confirm that, from my limited
experience and that of my friends, the following generally seem to work
well:
6.3
   all the standard Acorn applications
   Impression 2.17 (but see below)
   Poster 1.21
   Artworks (all versions)
   Vector
   TWO
   Squirrel
   Pacmania
   E-Type
   Twin World
   Lemmings
   Gods
6.3
Programs that need minor modifications
6.3
•   Sigmasheet (Minerva) must have FPEmulator in the same directory, even
though that module is in ROM. (John Waddell)
6.3
•   First Word Plus release 1.1 − As my current word processor is First
Word Plus release 1, I was worried by the assertion in the release notes
that this was unuseable under RISC-OS 3.1. This happily proved not to be
the case. If colour 14 within the WIMP palette is redefined, the menus
become legible. I have created an obey file to invoke 1WP (on the
monthly program disc) which loads a modified palette on entry and
restores the default palette on exit. Using this, I have experienced no
problems running 1WP. (Ed Harris)
6.3
•   DiscTree does not work correctly under RISC-OS 3. Oak Solutions are
working on the problem.
6.3
•   InterWord – My copy of InterWord has its own directory in the hard
disc root directory. It runs in 65Tube, which isn’t where it used to be
under RISC-OS 2, and appears under a different name. To get InterWord
going again, copy the new App2.!65Tube.!RunImage to the InterWord
directory and rename it as 65Tube; it should now be sitting alongside
!IWord in that directory. In the I/WORD.!IWORD.!RUN file, insert the
following line after Set Interword$Dir <Obey$Dir>:
6.3
RMLoad ADFS::HD4.$.IWORD.65Tube
6.3
(where HD4 is the name of the hard disc). Double click on !IWORD in the
usual way to start it.
6.3
A similar method should apply to single floppy machines by substituting
the appropriate floppy name for that of the hard disc above. (Ken Cowap)
6.3
•   Holed Out – I have a boot file on the hard disc which sets up the
machine with some applications on the iconbar and some more on the
pinboard. The golf game “Holed Out” finds this environment disagreeable
and won’t start. The following gets it going:
6.3
*CONFIGURE DRIVE 0
6.3
*MOUNT :0
6.3
   <shift-break> (machine re-boots to bare-bones iconbar)
6.3
   Click Drive 0 icon to get “Holed Out” filer
6.3
   Double click !Boot
6.3
Terminate the game with:
6.3
   Shift/Break
6.3
*CONFIGURE DRIVE 4
6.3
*MOUNT:4
6.3
   <ctrl-break> (machine re-boots to populated iconbar, etc) (Ken Cowap)
6.3
•   Chocks Away – The flight simulator Chocks Away isn’t quite so fussy
about a populated iconbar and will generally start normally. However, if
it does object, it may help to try a <ctrl-break> followed immediately
by <escape>, which re-boots to an unpopulated iconbar, then double click
on !Chocks. (Ken Cowap)
6.3
Other program problems
6.3
Mike Williams has supplied a list of programs that he has found don’t
work under RISC-OS 3.1:
6.3
   Ground Control Teletext Adaptor (At the A.U.      Show, Ground Control
said they weren’t      aware of any problems. HE.)
6.3
   Clares’ Arcade Games
   Quazer
   StarTrader
   Arcendium
   Corruption
   Terramex
   Boogie Buggy
6.3
Ed Harris has found several public domain applications that have failed
to run:
6.3
   Mandelplot (garbage on screen − machine      locks)
   ChartDraw (“illegal window handle” error)
   Hyper (“illegal window handle” error)
6.3
•   Computer Concepts admit that ShowPage does not work with RISC-OS 3 and
say that they have no plans to get it working with RISC-OS 3. There just
is not the demand for this product to make it worthwhile doing the
upgrade which is more like a complete re-write for the new operating
system.
6.3
Denis Fox reports problems with:
6.3
•   Impression (2.17) and LaserDirect (2.05) work much more slowly, even
with Fontsize set to 256Kb and Fontmax to 512Kb. There are regular poor
printouts, sometimes with patterned effects in what should be black
areas, hang ups when transferring lots of text between documents and
continual error messages when printing. (I have a similar setup but have
not noticed any deterioration in performance. HE.)
6.3
To help the speed, I suggest you try NOT using the ROM fonts − then you
can have Quick Text switched on. To do this, add a FontRemove
Resources:$.Fonts. in the !Fonts.!Boot file before the FontInstall
<Obey$Dir>. command.
6.3
I have once had the patterned effect in black areas but quitting and
restarting the printer driver cleared it. Ed
6.3
•   Multistore V11 not printing with a Star LC24-200. (At the AU Show,
Minerva said there may be problems with earlier versions and that users
experiencing such difficulties should contact them quoting their version
number. HE.)
6.3
•   Scanlight Plus 256 not working at all. (There are no known problems.
Are you sure you have a power supply connected to the board? You do need
the p.s.u. connected even on the A5000. Ed.)
6.3
•   A number of people have had problems (mainly “illegal window handle”
errors) with versions up to 1.07 of my PD application PickAPic. I found
that there was an error in the code that only occasionally caused
problems in RISC-OS 2, but caused the program to crash quite often in
RISC-OS 3, presumably because of a change in the way the window manager
numbers its windows. Anyone who wants an upgrade to the latest version
please write to me at the address at the bottom of this column. (Hugh
Eagle)
6.3
•   Droom – I can’t get Droom to work. It reports “Not enough system
sprite memory” whether I use Memalloc and the *Spritesize command (this
used to work in RISC-OS 2), or the new *ChangeDynamicArea command, or
whether I set the System sprites slider in the Task Manager window.
(Hugh Eagle)
6.3
•   Compression − Computer Concepts have informed me that the release note
for version 3.10 is slightly too pessimistic when it claims that
Compression (up to version 1.10) does not work on RISC_OS 3.10. What
apparently happens is that sometimes when you compress a directory with
several files, the last file, instead of being processed, produces an
error message. It can then be compressed singly. (In the first two weeks
of working with the new OS, I have had no trouble with Compression, so
the problem appears to be marginal. HE) (Jochen Konietzko)
6.3
•   Impression and the Pinboard − Can anyone tell me why the Pinboard has
such problems with Impression documents?
6.3
When I drag a compressed document onto the desktop, the next reset will
produce an error message “Sprite does not exist”. Without Compression,
the document does appear on the desktop but the Impression icon is
replaced by the standard application icon. (Jochen Konietzko) (I seem to
have no difficulty with Impression icons on the Pinboard in my
Compression-less setup. HE)
6.3
•   PC Emulator (version 1.6) – I applied the patch to my copy of the PC
Emulator (Version 1.6) and changed the filetype of the DOS partitions to
DOSDisc as recommended in the release notes. I then found that the
machine locked whenever I ran the emulator, even after a <ctrl-break>.
Things worked perfectly, however, after the machine had been powered off
and restarted. (Ed Harris)
6.3
•   Trouble with !FontFX − It has already been documented, at least twice,
that !FontFX will not run under RISC-OS 3 and the suggested “fix” is to
reset Font$Path using *SetMacro. Having done this, I still can’t get it
to use the Trinity and Homerton fonts supplied in ROM. To get round the
problem, I have renamed the original RISC-OS 2 versions OldTrinity and
OldHomertn and installed them in my !Fonts directory on disc. Now
choosing these from the fonts list in !FontFX works beautifully. So far
there have been no side effects!
6.3
It occurs to me that the problem may be that !FontFX is looking for
files called IntMetric and Outlines (there are lots of references to
these names in the !RunImage) whereas the files in ResourcesFS are
stored as IntMetric() and Outlines(). The real answer would be for the
program to be amended to accept either version of the filenames. I shall
speak to DataStore to find out if they intend to make the necessary
alterations. (Tom Harkness)
6.3
The release notes say that FontFX must be more recent than 4.31 to work.
My version, 4.08, works fine so long as you don’t use ROM fonts. If you
do try to, it reports ‘Can’t find outlines’ and promptly exits from the
iconbar! All other fonts seem fine, even from multiple directories.
(Steve Hutchinson)
6.3
•   Ovation – Ovation meets a problem or two with RISC-OS 3 despite its
recent update (version 1.35S). If text is indented (e.g. at the start of
a paragraph) and is also justified, the correct format is presented on
screen but when printed, the indented line projects to the right beyond
the text frame. It is understood that Risc Developments are working on
this. Also, if Ovation encounters a faulty outline font (e.g. containing
a zero-width character, not uncommon in some PD fonts) it puts out an
incorrect error message saying that the font cache is full. This bug is
more likely to be purely Ovation’s rather than in conjunction with RISC-
OS 3, but hasn’t been resolved as yet (except of course by ensuring that
outline fonts are fault-free). (Ken Cowap)
6.3
•   ArcDFS – Since fitting the RISC-OS 3 upgrade, ArcDFS no longer works
correctly and consistently comes up with error messages. This causes me
great problems since, throughout our group of companies, we possess some
50+ Acorn BBC units ranging from the old BBC Model B’s (yes, they’re
still going strong after eight years of service) through to Master
128’s.
6.3
We are only just moving on to Archimedes and to maintain our wealth of
information accumulated on our Beebs, I am transferring files from DFS
to ADFS via a set of external 5¼“ disc drives I have had connected to my
A3000. This has worked perfectly in the past with complete surfaces of
DFS discs being transferred to ADFS in seconds. Without ArcDFS, I am
left with only one solution − to *MOVE individual DFS files to 640Kb
ADFS discs using a Master Series Computer and then transferring these
640Kb discs to the Archimedes. Clearly, this is a retrograde step after
the convenience of ArcDFS.
6.3
I have been in contact with Dabs Press and, although they are aware of
the problem, they could not tell me whether or not they were going to
make the necessary modifications to their software! I sincerely hope
they receive a large enough response from their users to persuade them
to produce an upgrade for this product. (S. J. Furnell)
6.3
•   !Alarm – A couple of tricky little bugs exist in the new Alarm
application. I’ve already reported these to Risc User and trust they
will advise Acorn in turn. If the Repeating Alarm facility is used to
operate, say, every first Monday of every month, it malfunctions when
the next month comes up to December. It advances a year instead of a
month, going from November 1992 to December 1993.
6.3
Perhaps less important, but still a nuisance if you want to use it, is
the bug in the automatic BST to GMT facility. It refuses to accept
entries for two different years such as were needed at the recent change
on 25th October. Before that date, it should have been in order to enter
6.3
   BST starts on Sunday 28th March 1993 at 01:00
6.3
   BST ends on Sunday 25th October 1992 at 01:00
6.3
but Alarm invalidates such an entry by altering the “starts” date to be
earlier than the “ends” date. If one capitulates and enters the two real
changeover dates for 1992, (hoping that Alarm will calculate forward
into 1993) Alarm does indeed respond with a date in 1993 but only about
11 months ahead, on 1st March 1993. There is a fundamental difficulty
for the programmer here. The changeover date doesn’t, as far as I know,
follow a properly defined rule and tends to be arbitrary. I suspect that
the programmer intended to issue a reminder to the operator 11 months
after the last change so that the true date, known by then, could be
entered. However, the algorithm obviously needs re-examination. (Ken
Cowap)
6.3
Can readers (or the manufacturers) offer any advice on how to overcome
problems with these programs (or any others that have proved
troublesome)?
6.3
Hardware problems
6.3
•   Oak SCSI “Free Space” display – My Oak SCSI hard disc won’t use the
new ‘Free Space’ window but uses the old RISC-OS 2 ‘stop-everything-and
press-space’ window. The Release notes suggest that SCSIFiler will fix
this for the Acorn SCSI card, but it doesn’t appear to for my Oak
(v.1.16). Any thoughts? (Steve Hutchinson) (The solution to this was
explained in Archive 5.5 p40. Ed.)
6.3
•   Digitisers, SCSI and the RS232 port – I managed to get RISC-OS 3
fitted to my faithful A440 only to discover that my Technomatic
(Brainsoft) Multipod Professional digitiser not only does not work with
RISC-OS 3.10 but the computer will not work with the podule fitted.
Richard Brain says that he was not aware of a problem and there are
currently no plans for an upgrade, but he will have a look when he gets
time to see if an upgrade is possible. In the meantime, I have no video
digitiser, no stereo digitiser and no extra RS232 ports.
6.3
I mentioned the problem to Malcolm Banthorpe and he tells me that his
Pineapple digitiser is now “unwell” with RISC-OS 3.10. I wonder whether
there are some differences in the way that podules download their ROM
modules and initialise under RISC-OS 3.10?
6.3
I also hear rumours that people are experiencing interrupt problems with
SCSI interfaces causing data loss on the RS232 port (interrupts not
returning quickly enough to avoid loss of incoming data on the port).
Apparently, uncached Morley cards may be particularly prone but I have
no first hand experience. It certainly fits in with Malcolm Banthorpe’s
experience on his digitiser which seems to be suffering from an overdose
of interrupts. (Tim Nicholson)
6.3
Laser Direct RISC-OS 3 printer driver?
6.3
It seems barely credible, but a whole year after RISC-OS 3 was first
released, the Rolls Royce of Archimedes printers still lacks a RISC-OS 3
printer driver. For many purposes this doesn’t matter, but rotated fonts
seem a bit pointless if you can’t print them out, and users with more
than one printer would like to be able to access their LD using the
Print Manager.
6.3
Readers have been getting contradictory explanations from Computer
Concepts: one was told that they’ve been too busy writing Artworks,
while another was told that Acorn haven’t given them a printer driver
specification yet! (There’s an element of truth in both. Let’s just say
that it is a non-trivial task to write a printer driver to work with
Acorn’s Print Manager. Ed.)
6.3
Other printer problems
6.3
•   Epson MX100-III – My printer is an ageing Epson MX100-III. Using the
standard settings provided by the Acorn printer driver, I find thin
lines can disappear. e.g. the staves in a Rhapsody score. Using the
highest resolution setting (120 x 216 dpi dithered) produces an
excellent standard of output but 1.5 to 2 times slower than the RISC-OS
2 driver (at a resolution of 120 × 144 dpi). (Ed Harris)
6.3
•   Print buffering – I have an Archimedes 420/1 expanded to 4Mb memory,
with twin Cumana 5¼“ disc drives and recently equipped with ARM3 and
RISC-OS 3.1. Ever since I installed RISC-OS 3.1, I have had a problem
with printing. I use Impression most of the time but the same problem
arises with Draw, Acorn DTP, DrawPlus, etc, etc. I shall put the problem
in the form of a question: What is the point of a printer buffer that
does not release control of the machine back to the user until the print
job is finished?
6.3
I wrote to Acorn as follows: Whatever setting I use for PrinterBuffer
Size, it just does not seem to make any difference. I have 4Mb of memory
and use Impression a great deal. Sometimes I set the PrinterBufferSize
as high as 512Kb but the computer does not return control to the desktop
until the document is completely printed. Under RISC-OS 2, I used a
printer buffer program and found that I was able to work on a new
document within a minute or two of sending the previous one to the
printer provided the buffer was set high enough to accommodate all of
the data being sent to the printer. Now the machine is always tied up
until the printing is complete. How does one get the printer buffer to
work? I use *Configure PrinterBufferSize to set the size of the buffer.
Is there another command that I can use to make it operational and is it
possible to turn it off (*Configure PrinterBufferSize 0Kb sets it to the
default of 1Kb) and use one of the many printer buffer programs that are
available?
6.3
Acorn (Peter Dunn) replied: “Although the printer buffer does work in
that data is passed into the buffer, control is not released back to the
user until the print job is finished.”
6.3
I have tried RMKilling the BufferManager but that stops all data being
sent to the printer.
6.3
I have devised a work-around but it is a little long-winded:
6.3
1. Have two versions of the chosen printer driver on the iconbar and
make them active.
6.3
2. Set one (F) to send the data for the printer to a file on the hard
disc and set the other (P) to send the data to the parallel or serial
printer.
6.3
3. Select (F) by clicking on it and ‘print’ the document. A file called
Printout will appear on the hard disc.
6.3
4. Change the name of the file called Printout to, say, Printoutx.
6.3
5. Select (P) by clicking on it and drag Printoutx onto the Printer icon
that is yellow.
6.3
The file will now print in the background and control is instantly
returned to the user. Step 4 is not strictly necessary but is useful if
you are printing longish documents and wish to, as it were, stack them
ready for sending to the printer. If one does not change the name to
something unique, there is a danger of overprinting one file with
another. Instead of the hard disc one may, of course, use a floppy or
RAMFS.
6.3
This is quite a laborious work-around when one has a printer buffer
built in (so one would think) to RISC-OS 3.1. A better arrangement must
be possible and so I should be grateful for any hints or tips the
readers can give. (M. P. Ireland)
6.3
•   Epson FX-105 – The driver for the Epson FX-85 does not work with the
Epson FX-105 even though the printer manual suggests that they are
virtually identical apart from the carriage width. The printout is
always aligned along the left hand side even when it is supposed to be
centered, etc. Furthermore, large letters, being sent in several
separate lines, are often incorrectly aligned from top to bottom. When I
raised this in my letter to Acorn, they replied as follows:
6.3
“The Epson FX-105 does not support the ESC $ (remove leading white
space) options supported by the FX-85 but does however support NLQ text
printing. The decision by Epson not to support ESC $ in its newer
printers has only recently come to light when testing the Epson LX-100.
We have written a driver for the LX-100 which I am sure you will find
works correctly with your printer.”
6.3
A copy was sent to me and I can confirm that it does indeed work. (M. P.
Ireland)
6.3
•   Bug in !Printers − If you configure two printers of the same type
(e.g. a 300dpi HPLJ-2 and a 150dpi HPLJ-2) then delete the first one,
!Printers crashes with an address error. (Mike Williams)
6.3
•   LaserJetII printer driver and Qume Crystalprint Series II – My problem
concerns the LaserJet II printer driver. Quite simply, I always get an
extra blank sheet for every page printed. Also stray characters appear
in the left margin. The Qume CrystalPrint Series II printer is very
similar to that used initially by Computer Concepts for their Laser
Direct series, but mine isn’t a Laser Direct model.
6.3
In the early days of RISC-OS 2, I vaguely remember I had a similar
problem that was solved only when Acorn (fairly quickly) brought out
another printer driver.
6.3
In my efforts to solve the problem, I’ve tried lots of things – like
changing margins, etc, etc. I’ve had long discussions with Gerald
Fitton, too. From him, I know that others have plenty of problems,
though perhaps not quite like mine. The silly thing is that all was OK
with the RISC-OS 2 driver but, despite the impression given in Risc
User, neither of us thinks that RISC-OS 2 drivers can be used with RISC-
OS 3. Is this correct, I wonder? If they can be used, that would be a
good temporary solution. (John Jordan)
6.3
•   Getting the page length right – After quite a number of abortive
attempts to set the page length for 12 inch fanfold paper, I finally
found out that any changes I make become effective (at least on my
combination of machines) only after I have saved all changes, quit and
then reloaded the printer driver. (By the way, the correct setting for
my Fujitsu DL1100 printer, using the Epson LQ860 emulation is, for some
inscrutable reason, 12.309 inches, 72 lines). (Jochen Konietzko)
6.3
•   Panasonic KX-P1124 – Probably my most difficult problem was with the
printer drivers. The old RISC-OS 2 drivers just have to be abandoned −
they merely generate an error message. My printer happens to be a
Panasonic KX-P1124 (24-pin dot matrix, monochrome) but Acorn haven’t
seen fit to include any Panasonic printers at all. Though they blandly
say that one of the others will probably be near enough, in fact, it
requires a considerable amount of research to be confident that the
range of control codes is (a) sufficient to fully exploit the machine’s
capabilities and (b) doesn’t contain any alien codes to cause unwanted
results. At the end of the day (or several!) I reduced the probables to
two, the Epson LQ860 and the NEC P6+ (despite the fact that both
machines are colour printers). I had misgivings about the NEC which
includes, in its control code repertoire, some which started with 28
instead of the old familiar 27 (Escape). The matter was finally clinched
when I was able to question an Acorn representative at the Acorn User
Show. He told me that quite definitely I should use the Epson LQ860. So
this I do, and without any problems so far. I haven’t even edited out
the control codes for the colours − the driver is used, unmodified. (Ken
Cowap)
6.3
Where are the hare and the tortoise?
6.3
Can I get the hare and tortoise icons that came with my ARM3 back on the
iconbar? Or do I have to tell people that if they want E-Type to go at a
playable pace they have to press <F12> then type “Cache off” then press
<return> twice before running it?! (I just use
6.3
RMEnsure Arm3Icon 0.9 Run scsi::4 .$.<leafname>.!Arm3
6.3
 in my !boot file and... hare presto! Ed.)
6.3
Programmers’ Reference Manual
6.3
At the Acorn User Show, a “beta” copy of the new PRM was on sale in
photocopy form for £120. The final version will apparently be available
“in the first quarter” of 1993. It is about 50% bigger than the RISC-OS
2 manual and will fill six volumes. From my brief inspection, it looks
as if it contains quite a lot more explanatory material than its
predecessor. I was told that neither the price of the final version nor
details of the trade-in arrangements (if any!) have yet been fixed.
6.3
Various hints and tips:
6.3
•   Problem with Impression dongle − You can’t write to a printer when the
Impression dongle is fitted unless you have run the Impression appli
cation first. (Mike Williams)
6.3
•   Summer time blues − Some C programs (e.g. PVray) get the time wildly
wrong when *BST is configured. (Mike Williams)
6.3
•   Bug in !SciCalc − It gives wildly inaccurate results to the <±>
function for some values, e.g. if you enter “1<Exp>56” then press <1/x>
to give “10E−57” then press <±> you get the surprising result: “10E−58”.
If you press <±> again, you get “10E−41”. (Mike Williams)
6.3
•   Where did this file come from? − There are quite a few ways in which
<adjust> clicking can be very helpful. I’m sure you all know that
<adjust> double-clicking a directory opens it and closes the parent and
that using <adjust> to close the directory opens up the parent. You may
even know that clicking the close icon with <adjust> while holding down
<shift> actually opens the parent without closing the directory. What
you may not have noticed is that you can do the same with (some)
applications. With most applications, it seems, <adjust> double-clicking
a file loads the file and closes the directory from which it came. (I
think this is a function of the Filer and is therefore true for all
applications and files. HE.) Then, if you click in the close box with
<adjust>, holding <shift> at the same time, the directory from which the
file came appears on screen without the file disappearing from view.
With Impression, however, there is no need to hold <shift> as well −
just using <adjust> on the close box brings up the directory viewer in
which the file is saved.
6.3
Oh, and I’ve only just discovered that if you draw a window’s scroll bar
with <adjust>, you can move about in both x and y directions! (Paul
Beverley)
6.3
•   Rubber-banding − Those who have gone straight to RISC-OS 3 without
reading the manual (like Ed!) may not have noticed that you can use
rubber-banding to select a group of files. So instead of <adjust>
clicking every individual file you want to select, you can click
<select> in the space between two file icons and, as you drag, a red
bounding box appears which you can release when it is over the files you
want. Using <adjust> instead of <select> simply adds files to those
already selected but note that if you go over a file that is already
selected, it is then de-selected. (Paul Beverley)
6.3
•   SerialDeviceDriver module − On an A5000, never unplug the module
called “SerialDevice Driver”, as the following strange things will
happen. After a power off/on the machine will boot up but, during
booting, an error message “Error from Task Manager” will appear and
there is only a “Cancel” button to click on. Then the boot up is
finished correctly but no Taskmanager is present! It will only reappear
if you double-click on any file with the new Desktop filetype (&FEA),
even if this is 0 bytes long! In addition, the Taskmanager will not
function correctly in all cases: try to get an “!Edit Task Window” now.
Bug or not? (Carsten Bussman)
6.3
•   Double-clicking into Basic − On my first machine (which was killed by
my dealer while trying to fit a SCSI hard drive) I was able to enter BBC
Basic V directly by double-clicking into the system’s start-up message
window (“RISC-OS ... initialising”). This does not work with my current
machine (Serial no. 1009939); I don’t know why. This was a nice feature
which I’m missing now, so why isn’t it implemented in all RISC-OS 3
ROMs? (Carsten Bussman)
6.3
•   Inserting path names into !Edit − Try dragging any file or directory
into an !Edit window while pressing the Shift key. This will give the
file’s pathname. (Carsten Bussman)
6.3
•   “Menu Area Full” error in FormEd – Captain K. J. Tompkins reports the
error “Menu Area Full (Internal code 6560,1)” from his version of
!FormEd. This is probably because he has one of the earlier versions of
this program which only allows limited space for building its menus and
is defeated if there are too many fonts installed. If so, it is fairly
easy to cure: in the Basic program called !RunImage inside the !FormEd
directory there is a line that reads something like:
6.3
DIM menufree% &1800: menuend% = menufree%+&1800
6.3
(this is at line 790 in the version I’m now looking at). If each of the
two &1800’s is replaced by &2000, so that the line reads:
6.3
DIM menufree% &2000: menuend% = menufree%+&2000
6.3
 the program should work.
6.3
(Note: this hint does not apply to all versions of FormEd) (Hugh Eagle)
6.3
•   Filer_Run syntax − it seems that the command
6.3
*Filer_Run <App$Dir>
6.3
(where App$Dir is a system variable set to the name of an application)
doesn’t run the application as you might expect. What it does is to open
a filer window on the application’s directory with “<App$Dir>” (not the
actual pathname) in the title bar.
6.3
The following syntax, however, does seem to work:
6.3
*Set Alias$FR Filer_Run <App$Dir>
6.3
*FR
6.3
*UnSet Alias$FR
6.3
(this is equivalent to the old trick that used to be needed in RISC-OS 2
to open a directory viewer with *Filer_OpenDir). (Hugh Eagle)
6.3
•   *FontRemove syntax − The FontRemove star command seems to be very
fussy. You have to tell it the name of the directory that you want to
remove in exactly the form in which it is included in the Font$Path
system variable. For instance, if you want to remove
RAM::RamDisc0.$.!Fonts. from Font$Path you must type
6.3
*FontRemove RAM::RamDisc0.$. !Fonts.
6.3
including both the disc name (RamDisc0) and the full stop at the end.
(Robert Chrismas has sent in a handy application called !FontsCut which
lists the directories currently included in Font$Path with numbers
beside them and all you have to do is type in the number corresponding
to the directory you want to remove. This is included on the monthly
program disc.) (Hugh Eagle)
6.3
•   Configuring Shift Caps − In the Caps lock section of the Keyboard
window of !Configure, the effects of “Caps lock on” and “Shift caps”
seem to be reversed. In other words, selecting “Caps lock on” means that
holding down <shift> reverts to lower case whereas selecting “Shift
caps” means that you get upper case regardless of whether <shift> is
held down or not. (Hugh Eagle)
6.3
•   Mouse button codes – If you enter and run the following simple Basic
program:
6.3
REPEAT
  MOUSE X,Y,B
  PRINT TAB(10,10)B
UNTIL FALSE
6.3
then watch what number is displayed on the screen as you press the
various mouse buttons you will see:
6.3
no button   0
6.3
Select      4
6.3
Menu        2 for an instant, followed by 0
6.3
Adjust      1
6.3
The behaviour of the menu button has changed from RISC-OS 2, I think.
Those programs that are interested only in detecting whether the button
has been clicked won’t be affected, but the change can cause problems
for programs that want to detect whether the Menu button remains held
down. Luckily, there is a simple solution: the function
6.3
INKEY(-11)
6.3
will return the value TRUE however long the Menu button is held. (Hugh
Eagle)
6.3
•   Getting Art Nouveau to run – Hands up those who remember Art Nouveau?
Well, this excellent old program has one quirk, namely that you have to
hold down the menu button while you move round the menu tree until you
make your selection. So, because of the changed functioning of the mouse
button (see above), it doesn’t work in RISC-OS 3.1 unless you make a
slight modification. In the Basic program called ArtNouveau from the
directory ANfiles at line 30950 in my version there is a line that
reads:
6.3
WHILE stayopen AND mbut=2
6.3
I have found that altering this to:
6.3
WHILE stayopen AND INKEY-11
6.3
restores the program to health. (Hugh Eagle)
6.3
•   Using a TaskObey file to display a message – The new filetype
“TaskObey” is useful if you want the !Run file of an application to pop
up a message in a window without halting the whole system. You might use
this for instance to display a warning to change the setting of your
printer. Thus, to display “This is a message” in a window, create a file
called, say, Message containing:
6.3
Echo This is a message
6.3
set the filetype to TaskObey and include, in the application’s !Run
file, the line
6.3
Filer_Run <Obey$Dir>.Message
6.3
If you had set the filetype to Obey, the message would pop up moment
arily then disappear before the application loaded. (Hugh Eagle)
6.3
•   Deskboot files and the Pinboard – The User Guide explains how to
create a !Boot file to boot the computer in your desired configuration.
It adds that, when saving the !Boot file, information to start up
applications produced before RISC-OS 3 is not saved automatically. To
get round this, it explains how to rename the !Boot file as Deskboot and
to incorporate it in a !Boot application. It goes on to explain how to
edit the Deskboot file, either to remove or to add lines as necessary.
6.3
If you use the Pinboard as I do, some icons will be changed on a regular
basis. For example, I have regularly used files which change each month.
If the Deskboot file is saved each time the Pinboard is changed, a
certain amount of what the User Guide calls “hand editing” would have to
be done each time a change is made, to remove or add relevant
instructions.
6.3
There is an easy solution. Set up your Desktop once and save the
Deskboot file. Then click <menu> over an empty space on the desktop and
save the Pinboard settings in the Pinboard file in your root directory.
Now drag your Deskboot file into !Edit (or Shift-double-click on it) and
delete all the lines which begin “Pin”. Alter the line “Pinboard” or
“Pinboard -Grid” to read “Run Pinboard” or “Run Pinboard -Grid”.
6.3
Future changes to the Pinboard setting now need be saved only to the
Pinboard file in your root directory. (John Waddell)
6.3
Ken Cowap warns that a DeskBoot file can easily be unnecessarily
cluttered if, for instance, there are a number of unwanted items on the
screen when it is made. This will incur a penalty in the time taken to
boot up − his started at two and a half minutes before he trimmed it
ruthlessly.
6.3
•   ADFS buffers − According to Acorn’s telephone help line, the bug which
made it advisable to configure the adfs buffers to zero has been fixed!
(According to Computer Concepts, unless the buffers are configured to
zero, it is vital to dismount disks before removing them for good from
the machine, because otherwise there may be trouble later.) (Jochen
Konietzko)
6.3
•   OS_Byte calls and the serial port – I use my A3000 for many purposes,
one of the more important being as a terminal to our mainframe computer.
The software for this I have written myself in Basic. After the upgrade,
my terminal program wouldn’t work. It would not accept characters from
the serial port. I tried using the serial port utility module supplied
with the upgrade, but this did not help.
6.3
After a lot of experimenting, we finally discovered that if we changed
all the *FX calls for SYS “OS_Byte” calls, then the program would work.
It still puzzles me why two supposedly similar commands should give
different results! (S. J. Furnell)
6.3
•   Z88 Filers – The Release Note says that the !Z88 filer supplied with
Pipedream requires interactive filing to be off. You’ll be pleased to
know that David Holden’s !Z88 filer (Careware 7) suffers from no such
problems. (Steve Hutchinson)
6.3
•   Programs that call machine code from Basic – An application I used a
lot, !BPrint (Risc User) wouldn’t run with RISC-OS 3, and the cure for
this may well apply to other applications organised in a similar layout,
i.e. using a central Basic program calling a separate piece of machine
code. The cure was to re-assemble the code under RISC-OS 3; obviously
there is some small difference between its assembler and that of RISC-OS
2. In this particular application, its “ancestor” (the program developed
over a period of time), a program called FileList from vol.2 issue 5 was
rerun to produce the assembled file BLIST which was substituted for the
original. (Ken Cowap)
6.3
•   Mode number conflicts – A program for plotting Mandelbrot diagrams
failed with RISC-OS 3 and the manner of its failure may indicate a cure
for others. This particular program called a Mode 40 which didn’t exist
under RISC-OS 2 but which was specially designed and provided with the
program. RISC-OS 3 of course does have a native Mode 40 and a clash
between it and the “special” gave rise to the error report. The cure was
to rename the “special” to a mode number outside the range provided with
RISC-OS 3, in this case Mode 50. (Ken Cowap)
6.3
•   Setting !Alarm options – This is a minor matter, unless you are trying
to set your own !Alarm display format in a !Boot file. This you do with
a Set Alarm$Options line (p136 App Guide). The quotes are omitted. This
is my (working) line:
6.3
Set Alarm$Options -format “%z24:%mi %zdy.%zmn.%zyr”
6.3
If you use a !Boot file and !Alarm, put it in to see my favourite
display format! (Steve Hutchinson)
6.3
Miscellaneous comments
6.3
•   Surely a ‘grown up’ operating system shouldn’t say, “Disc error 10 at
:0/00000000” when you try to read an unformatted disc?
6.3
•   It seems strange that !Configure requires you to press <return> after
typing into a writable icon. This is not normal RISC-OS convention.
6.3
•   It’s a shame that they didn’t manage to squeeze some suitable backdrop
sprites onto one of the RISC-OS 3 discs. Those who have access to
Windows 3 might like to know that its wallpaper bitmaps seem to work
very well on the RISC-OS 3 Pinboard, and it is very easy to transfer
them from an MS-DOS floppy, via !ChangeFSI (on the Support disc), into
sprite format on an Archimedes disc. The file called “leaves.bmp” makes
an excellent seasonal backdrop (I’m writing this in October!).
6.3
Alt-key combinations
6.3
The last couple of issues of Archive have contained lists of Alt-Key
combinations in RISC-OS 3.1. To complete the picture I would add:
6.3
To get accents ( ´ ` ¨ ^ ° ¸ ) on their own, press the appropriate Alt-
key combination (e.g. Alt-[ for a ´ ) then the space bar.
6.3
Alt-Spacebar gives a “hard space” or “non-breaking space” (the sort you
put between “NR2” and “2SD” if you don’t want Archive’s postcode
“NR2 2SD” to be split over two lines.)
6.3
I have put a diagram of the keyboard showing the Alt-key combinations
below, and I hope that Paul will put the drawfile on the monthly program
disc.
6.3
Incidentally, I am a bit puzzled about the character (code 173) that is
produced by the Alt-hyphen combination. What’s it for? I thought at
first that it was the “soft hyphen” that is described on page 109 of the
Impression 2 manual, but I now see that this is produced by Ctrl-hyphen
(this character – code 31 – is normally invisible but indicates a
preferred place for breaking a word with a hyphen at the end of a line;
it works even if the hyphenation module is not present). By contrast,
character 173 is, so far as I can see, always visible (as a hyphen if
the font follows the Acorn standard) but its effect is to prevent a word
into which it is inserted ever being broken over the end of a line even
if hyphenation is switched on. Am I correct in describing it as a “hard
hyphen”? (Well, I can think of one application. Suppose you want to use
a word like “pray­ers” where you deliberately hyphenate the word to
refer to the people doing the praying. In that case, if it were an
ordinary hyphen, the word might get split at the end of a line. Using a
‘hard hyphen’ avoids the split. Ed)
6.3
The upgrade again
6.3
Finally, I thought it would be good to take a slightly light-hearted
look at the subject as we hear the experiences that Rob Wears had in
fitting and using RISC-OS 3 on an A310...
6.3
I finally managed to get my grubby mitts on the RISC-OS 3 upgrade at the
Acorn User show, and arrived back home with that odd mixture of terror
and excitement that comes from having to take the lid off the machine.
I’ve fitted one or two things in the past, like the soldered 4MByte RAM
upgrade, but no matter how many times I do it, it’s still a bit like
performing neurosurgery on a close friend. Will Archie pull through,
Doc? Actually, fitting was a lot easier than I’d thought. I used the
IFEL carrier board, as their memory upgrade has proven so reliable. My
only source of concern was the discrepancy between their fitting
instructions and the ones supplied by Acorn which alluded to changing
links on the mother board. I followed the IFEL guidelines, which had the
link changes on their board, and waited anxiously for the machine to
explode when I switched it on. No problem, apart from the fact that I
bypassed the desktop and ended up at the command line. Still, the
desktop was easily accessible with a simple “*” command.
6.3
The hard drive icon was where it should be (and no longer jumped around
when I mounted or dismounted the disc!) and I am pleased to say that I
encountered no problems with the Oak SCSI board (version 1.16 of SCSI
driver).
6.3
I couldn’t resist the urge to delve into the Apps directory, and here I
encountered my first problem. Initially, I couldn’t get !Edit to load
from Apps and was greeted with the wonderfully informative message ‘File
“Basic” not found’. Fighting the sudden tide of panic (“It wouldn’t work
at all if the ROM legs were snapped off, idiot!”) I checked the module
list via the command line and found that I was apparently the lucky
owner of the one set of chips which Acorn had shipped without putting
Basic on board first. I was about to fall on my soldering iron in
despair when I thought of checking the status of the system modules
using *ROMModules and found that the Basic module was unplugged. I
reinitialised the module, and !Edit worked fine. There, that wasn’t too
bad − the hair will probably grow back and nobody will notice the
teethmarks on the manual....
6.3
Incidentally, the ARM3Support module was also unplugged and therefore
the *Cache command wouldn’t work, but I’d got the hang of things by now,
and pressed on happily. I then UNPLUGged the modules that I thought I
wouldn’t need (old habits die hard) like the Econet and NetFS modules −
each unplug command was greeted with the comment “xxxx module not found”
yet a subsequent check revealed that the module had been unplugged
correctly − odd! Archie hadn’t behaved like this before his lobotomy. It
was a bit unnerving, like suddenly finding the Queen Mother skateboard
ing down your front drive.
6.3
I then set about gingerly reconstructing the !Boot option for the hard
disc, feeding the bits to my new Archie in small, digestible chunks.
Compression (version 1.10) and !Spark both worked fine, after I
remembered not to set the “Newer” option for copying. However, Archie no
longer likes the “OS_UpdateMEMC” call which I had used to speed up the
ROMs. This is now treated with the same degree of horror as if I’d just
offered it deep-fried weevil for tea, and the machine has to be kicked
out of its sulk with a hard reset. I initially and unjustly blamed the
old OSSys module, but this actually seems to be fine and is the only
real survivor from my old collection of useful bits and pieces − have
they STILL forgotten to put a *SYS command in?
6.3
The next little gem was the “ChangeDynamic Area” calls which are
automatically saved in the desktop !Boot file − these override the CMOS
RAM settings and I couldn’t work out why the system sprite area remained
so large despite all my efforts to shrink it. And remember, Gentle
Reader, not to save the !Boot file in a Compression directory because,
despite what they all say, those teethmarks in the manual do show! On
the brighter side, I found that the configured startup language was not
the desktop, which explained Archie’s new enthusiasm for the command
line. The number for the configured language of your choice is now
obtained by counting down the list of modules − how simple. Not a
mention of alchemy or the Black Arts. Good grief, if it’s that simple,
anybody will be able to do it...
6.3
The !FontsPlus application caused some frontal headaches (actually,
these may have been caused by sudden and repeated violent contact
between the forehead and the monitor) as I could not persuade it to
recognise the ROM based fonts. Off it went to the great filestore in the
sky, and back I went to using multiple directories with original names
(!Fonts, !Fonts2, !Fonts3, !Fonts4, etc.). How quaint!
6.3
On to some serious computing. Impression works! (Actually, I cheated. I
got the upgrade from those nice CC people at the show). And there are
more screen modes available from startup, and a *WimpMode command to
simplify my DTP !Boot sequence. On the down side, I have found printing
to be a bit of a pain − my old !PBuffer application doesn’t work anymore
and even making the PrinterBufferSize HUGE using *CONFIGURE doesn’t seem
to do anything. However, there are many more options available for
printed output. I can even get 360×360 dpi on the Panasonic and it only
takes about 30 minutes a page!
6.3
I seem to have more space on the hard disc now. I used to have lots of
little applications that I’d collected over the years which now have
their functions in ROM, such as SetType, FileFind, StickyBD and
ShutDown. Being ruthless, and having no more need for them, I erased
them completely. Who knows, one day, I might even erase the backups of
the backups.
6.3
So am I sorry to have upgraded? Not at all. It was just that the scale
of the change surprised me. I suppose that the change was more compli
cated because I had my system set up exactly the way I wanted it, and
had to restructure virtually the whole hard disc. It was just as
traumatic as upgrading from Arthur to RISC-OS2 had been, but I’m very
pleased with my new machine. Of course, 3.1 goes into 310, and it may
even be 100 times better!
6.3
(Actually, this has an important moral − if Rob had followed Acorn’s
instructions and done a <delete-power-up> in the first place, he
wouldn’t have had half the problems he did. Ed.)
6.3
And finally ...
6.3
Send your comments, suggestions, queries, complaints, hints and tips
either to me, Hugh Eagle, at 48 Smithbarn, Horsham, Sussex  RH13 6DX.  A
6.3

Hints and Tips
6.4
•   Refilling BJ300/330 Ink Cartridges − The cartridges have a “soak up”
pad in them as well as the ink sachet so if the pad is black and full of
waste ink you can’t refill them! or can you? This is how to do it.
6.4
Split the cartridge along the seam and lift off the top. It is only
lightly tacked together so its quite easy to prise apart.
6.4
(The next bit is messy so use disposable gloves so you do not get ink
over your hands.)
6.4
Lift out the ink sachet (keep it the same way up) and then the pad and
wipe clean the cartridge. Dispose of the pad in a plastic bag, trying
not to get too messy.
6.4
I used kitchen roll for the new pad. Cut about 12 pieces to the size and
shape of the pad then place in the cartridge, now place the sachet on
top. Keep it the same way round to make sure the needle hole lines up
with the needle in the printer.
6.4
That’s what the safety flap is for that drops down when you remove the
cartridge so you don’t jab yourself.
6.4
Replace the top of the cartridge, securing it with just a couple of
drops of glue so it can be removed again, it is now ready to be filled.
6.4
I took the advice of Stuart Bell (Archive 4.5 page 7) and used Quink
Permanent Black @ £1.70p for 54ml. The sachet takes 40ml, so don’t over-
fill it. Obtain a 20ml syringe and 21 gauge needle from your chemist. (A
20ml syringe is about the best as larger ones tend to be very
expensive.)
6.4
Carefully fill the sachet making sure the needle goes in the same hole.
If you make another hole, air can get in and cause problems.
6.4
Don’t forget to wash out the syringe and needle afterwards.
6.4
So, at the cost of approx £1.70p plus the syringe and a few sheets of
kitchen roll, you now have a refilled ink cartridge which would normally
cost about £10 to refill and between £15 − £20 for a new one.
6.4
If any of you wonder what the soak up pad is for, when the printer is
switched on, ink is jettisoned through the nozzles to clean them and
this then runs down into the cartridge.
6.4
•   D-type connector problems − In Archive 4.8 p7, there was a tip about
taking the washers from under the hexagonal pillars of the D-type
connectors. I think it is worth re-stating it for those readers who
missed it. My A310 was in a ‘not working’ state the other day − it
refused to produce any green. I checked that the two screws holding the
D-plug into the socket were tight and checked out the monitor and lead
which worked fine, but still no green. A few days later, I remembered
the tip, removed the washers and the green returned.  Les May,
Rochdale

Using RISC-OS 3
6.4
Hugh Eagle
6.4
I hope that all who have ordered RISC-OS 3 upgrades will have received
them by the time this issue is published and that those who have been
waiting to make up their minds about upgrading will have been suffi
ciently encouraged by now to go ahead. There will be considerable
advantages to all Archimedes users if we all use the same operating
system: above all programmers will be able to use the new improved
features without having to worry about backwards compatibility.
Inevitably, this column will tend to contain quite a lot of negative
material about difficulties that some people have had, but it is
important to appreciate that most upgraders have had very little real
trouble (some inconvenience, perhaps, but not insuperable problems).
There are genuine advantages in RISC-OS 3, most programs work well and I
believe that most users are pleased with it.
6.4
The deadlines for the December and January editions of Archive have been
concertina’d close together (because of Christmas, etc) so there has
been no time for any reactions to the first RISC-OS 3 column. However,
contributions have continued to arrive in response to the original
request – several more problems, criticisms, etc but also some interest
ing hints & tips.
6.4
Please send anything that you think may be relevant either to Archive or
to me, Hugh Eagle at 48 Smithbarn, Horsham, Sussex  RH13 6DX.
6.4
Fitting the chips to an A310
6.4
The upgrade to my A310 with an IFEL carrier board was quite straightfor
ward, with the help of good instructions. IFEL recommend that you remove
the motherboard, as a fairly good pressure is required to insert the
carrier board. I encountered no problems but the soldering of three
wires from the carrier board on to the legs of a chip, is not for the
faint hearted. (Memories of my Beeb days came flooding back). (Roger
Power.)
6.4
Configuring for a SCSI hard drive
6.4
The configure application is OK apart from the settings appertaining to
my Oak SCSI filing system. I have overcome this by a separate !Config
Obey file that I run as well as Configure and which reads as follows:
6.4
*configure drive 4
6.4
*configure filesystem SCSI
6.4
*SCSI
6.4
*configure SCSIDrive 4
6.4
*OPT 4,2
6.4
This is obviously only needed after a factory reset which shouldn’t be
very often. Unfortunately, at the moment it is, as further problems have
now occurred. Namely my Impression doesn’t read the Dongle all the time
and appears to require a factory reset. This is only since the arrival
of RISC-OS 3.1 and I have yet to find out whether or not the Dongle (or
maybe even RISC-OS 3.1) is at fault. Obviously I have checked all the
connections over and over again. (Roger Power.)
6.4
(Roger, I suspect it’s more likely to be a hardware problem on the
p.c.b. after fitting the I.F.E.L. ROM upgrade in your A310. There are no
problems, that we know of, with dongles. Ed.)
6.4
5¼“ floppy drive
6.4
My external 5¼“ floppy drive is too slow for my Archimedes, with the
result of the heads chattering alarmingly. I had previously got round
this problem with the command *Configure STEP 3 6. However, RISC-OS 3.1
does not appear to accept this, only STEP 3 3. Can anyone help, please?
(Roger Power.)
6.4
LaserDirect printer driver
6.4
(Incidentally, Computer Concepts have told me that “RISC-OS 3 compatible
printer drivers should be available around the middle of next year. This
will cover TurboDrivers, FaxPack and LaserDirect” HE)
6.4
In RISC-OS 2, I solved the problems of using a LaserDirect printer on a
machine to which a parallel printer is also attached, by including a
simple driver in an application which resides on the iconbar and does
lots of idiosyncratic things which suit my personal use of the machine.
The driver stores data dragged to it in a buffer and outputs it to the
dot-matrix line by line on each Wimp_Poll with reason code 0. This
enables me to leave the LaserDirect on the iconbar. If I want to print
some text or Basic, the data is dragged with <shift> held down,
otherwise the data is presumed to represent an address, which is
formatted for the 9-line labels with which, by default, the printer is
loaded. This is very handy, because I can drag the address directly out
of a DeskEdit or Impression window. My beef is that my routine orig
inally tested for a connected printer, thus:
6.4
DEFPROCChPr:*FX5,1
6.4
VDU2,1,0,3
6.4
SYS“OS_Byte”,152,3 TO ;f%: f%=f%AND2
6.4
IFf%=0SYS“OS_Byte”,145,3
6.4
*FX5,5
6.4
ENDPROC
6.4
Thus, if the 0 byte inserted into the printer buffer is still there, the
printer is not on and the data are output to screen via a command
window; otherwise, they go to the printer; f% being the determining
flag. Now, I can understand that it is an enhancement to RISC-OS to wait
if the printer is off, but I find it is wholly unacceptable that the
machine is locked up by so simple an event. The OS_Byte no longer works,
since presumably the machine is stuck at the end of line 3200! Now, the
RISC-OS driver is not stopped by such an event, but multitasks happily
waiting for the printer to be switched on. Clearly some routine is
available to deal with the situation; why could the release documenta
tion not give a clue?
6.4
I have had to abandon the routine temporarily because, if I wished to
display a file on screen, the printer was incidentally tested first.
This is no longer possible, unless I validate <escape> with *FX229
before calling the routine and reset on exit; this means exiting via the
error handler, however, and is one of the more difficult parts of Wimp
programming, I find! Any suggestions? (Mick Day.)
6.4
LaserDirect and ROM fonts
6.4
As mentioned last month, the current version of the Computer Concepts’
LaserDirect printer drivers are not fully RISC-OS 3 compatible. Thus
they will not cope with the extra features of Draw such as rotated text
or sprites. Furthermore, they will not find all of the fonts in the ROM-
based resources filing system. In particular, the italic versions of
both Homerton and Corpus are not seen. It is necessary to have the old
copies of the Homerton and Corpus fonts installed in the disc !Fonts
directory. (Brian Cowan.)
6.4
(I find that if I try to print these fonts from Impression, it refuses
to print but does allow me to carry on, whereas Draw locks up the
machine. If this is a printer driver problem it makes Computer Concepts’
apparent lack of interest in bringing out new drivers all the more
frustrating. If they do appear in “the middle of next year” that will be
about 20 months after the original release of RISC-OS 3! HE)
6.4
Dot matrix printers: formfeeds
6.4
I have yet to find a way of stopping the RISC-OS dot-matrix printer
driver from issuing a concluding formfeed when a short piece of text is
dropped on the icon and printed. I prefer to control such things myself;
any ideas? or am I, after all, an idiot? (Mick Day.)
6.4
DFS discs: don’t try “free” space
6.4
Put a DFS disc in drive 0 and click; the error window tells you it is an
unrecognised format. However, if you put a DFS disc in drive 0 and click
on ‘Free’ you will crash the machine with ‘File core in use’ rearing its
ugly head! Couldn’t the existing disc checking routines have been
invoked before plunging into the ‘Free’ sequence? (Mick Day.)
6.4
Naming the RAM disc
6.4
RISC-OS 2 used RAM:$ as the prefix on the RAMdisc filer window and
Filer_OpenDir worked with it quite happily. However, you had to find out
by trial and error what discname to use in order to get sense out of,
e.g.
6.4
SYS “RamFS_FreeSpace”,“RamDisc0” TO room% : IFroom%>X% room%=0 : ENDPROC
6.4
Acorn could have detailed this irritating time-waster somewhere (like
the PRM). Now, I appreciate that RAM::RamDisc0.$ is generically the
correct format to use in the Acorn file system protocols and clicking on
the icon in RISC-OS 3 opens a window with that title. For backwards
compatibility, *Filer_OpenDir RAM:$ still works; jolly good. However, if
your software issues the command and you have already opened a window,
RAM::RamDisc0.$, from the iconbar, you get a second window identical in
all save the title, RAM:$! (Mick Day.)
6.4
Operating system version
6.4
The PRM tells us that SYS“XOS_Byte”,0,0 prints the version string: so it
does. However, SYS“OS_Byte”,0,1 TO ,A is supposed to return the OS
version number in A but it does not seem to do so. RISC-OS 2 and RISC-OS
3 on my machines both return 6 (the SYS number for OS_Byte)! However,
SYS“XOS_Byte”,0,0 TO A does return, in (A+4), the address of the version
string, so it can be extracted in the time-honoured manner. (Mick Day.)
6.4
MS-DOS: file extensions and formatting
6.4
The MultiFS utility is not required in RISC-OS 3; DOS discs and DOS
partitions on hard discs may be read directly with the new operating
system. I think that what Acorn have done is to integrate MultiFS into
the machine’s filers as an Image filing system. However, I cannot obtain
all the old facilities of MultiFS. In particular, it is not clear
whether the hierarchical method of dealing with file extensions can be
used; this was particularly useful. Also, I see that you can copy the
boot sectors from one DOS disc to another from the command line using
*CopyBoot, but I don’t know if this can be done from the desktop as an
option while formatting. Does anyone know? (Brian Cowan.)
6.4
Compression
6.4
CC have told me that there is a bug in version 1.10 of Compression which
can cause a corruption if a file is dragged between a CFS window and the
corresponding uncompressed filer window. If this is not done, Compres
sion should work OK. (Brian Cowan.) (See also the comment by Jochen
Konietzko in last month’s column.)
6.4
Wish list for RISC-OS 4 (R. W. Darlington)
6.4
•   When I use Name Disc from desktop, I want to see it display the
present disc name in the writable icon. (With my setup it does display
the name of the hard disc but not that of a floppy! HE)
6.4
•   When I copy across a group of files from one directory to another, I
want to see the Filer check that there is sufficient disc space for the
copied files to fit before it starts to copy them. (Also, in the case of
D format discs, that the disc won’t need compacting.)
6.4
•   I want to see it copy a group of files from one disc to another a
little more intelligently than it does at the moment. I wish to see it
write all the directory information in one go, then write all the files,
instead of continually moving the heads across the surface of the disc
for each little file it writes to disc, which takes it an age!
6.4
•   When I Set Type from Desktop, I want to see a list of possible
options, along with their icons, like !SetType by Emmet Spier.
6.4
•   In !Edit, I want to see an option to ‘Select All’.
6.4
•   In !SciCalc, which uses BASIC64, I want to see it display figures to
the full precision of the Basic, 18 significant figures or so and not
the 10 displayed at present. Also, I want to see it display exponents
like this “×1018” and NOT “E18”. I also want to see an option to select
engineering or scientific notation.
6.4
•   When deleting columns at the far right hand edge of a sprite, it
deletes too many columns.
6.4
•   If I save a sprite (using !Paint) of just one pixel in height (any
length?) and colour it a different colour than desktop grey, then when I
use this sprite to create a backdrop using the !Pinboard, not only does
it take an incredible amount of time to draw the backdrop, it also uses
the wrong colour. It comes out in desktop grey! Is this a problem with
!Paint or !Pinboard? (Or is the user making unreasonable demands? HE)
6.4
•   In the Set Copy Options, I want to see an option whereby it can be
made to ask for confirmation only if a deletion is being made, rather
than just an access request or a copy command.
6.4
•   Now that some fonts are inside RISC-OS 3 itself, I want to see a
configure option to set the desktop default font to other than the
system font.
6.4
•   I also want to see a new MODE equivalent to MODE 31 in every way
except that the pixel units are set such that all icons and everything
on screen appears half the size. Or, do the same for MODEs 18, 19, 20,
and 21.
6.4
•   When I insert a disc with a foreign format, e.g. an IBM formatted
disc, into the drives, I want to see a little more obviously that it is
a foreign format. Why not have the window header in a different colour,
e.g. red! Or do it some other way if you must, but make it immediately
obvious that it is not a native format disc.
6.4
•   When a window now goes off screen and the bottom right-hand ‘size’
control is grabbed, the window increases in size upwards when it should
not. (Isn’t this one of the intended improvements in RISC-OS 3? HE)
6.4
Pinboard backdrop sprites
6.4
R. W. Darlington has also sent in a collection of 74 sprites suitable
for tiling the backdrop, a voice module (which sounds to me as if it
might have been generated not a billion light years from the Sirius
Cybernetics Corporation) and some suggested modifications to the !Boot
file which will sound the voice module and display a different backdrop
each time the computer is re-booted.
6.4
The modifications to the !Boot file read as follows:
6.4
RMEnsure PinVoice 0 RMLoad PinVoice
6.4
CHANNELVOICE 1 10
6.4
SOUND 1 &FFF1 60 160
6.4
WimpMode 31 :REM to suite my sprites
6.4
Run ADFS::HD4.$.PinSp.RenSprites
6.4
Pinboard
6.4
Backdrop -T ADFS::HD4.$.PinSp .Sprit01
6.4
This assumes that the PinSp directory contains the backdrop sprites and
an Obey file called RenSprites which simply cycles the sprite names and
reads as follows:
6.4
Rename ADFS::HD4.$.PinSp.Sprit01 ADFS::HD4.$.PinSp.Sprit00
6.4
Rename ADFS::HD4.$.PinSp.Sprit02 ADFS::HD4.$.PinSp.Sprit01
6.4
......
6.4
Rename ADFS::HD4.$.PinSp.Sprit00 ADFS::HD4.$.PinSp.Sprit74
6.4
Screen blanker configuration
6.4
Does anyone know which *configuration command sets the blank out time,
or where this appears in the *Status data. I can set it using Acorns
!Configure program in the Apps directory, but cannot find it elsewhere
for use by a star command. (Roger Darlington.)
6.4
*Filer_Run syntax
6.4
My tip last month about using *Filer_Run with a variable application
name was unnecessarily convoluted. Although *Filer_Run <App$Dir> doesn’t
work, *Filer_Run <App$Dir>.!Run does.
6.4
Fitting RISC-OS 3.1 to Computerware/Avie/Atomwide RAM upgrades
6.4
The new RISC-OS ROMs are easily fitted into the four sockets provided on
the main RAM board. The ROMs should be fitted as described in the
instructions supplied with the RISC-OS upgrade.
6.4
Locate the two links on the RAM board and swap the jumper to position B
for both.
6.4
Locate LK12 on the main PCB, this is just under the RAM PCB where the
cable enters from the left. There should be two links placed in the
east-west postion. Remove these links and replace them in the north-
south position.
6.4
LK12 is not always fitted and exists as a pair of PCB tracks on the
topside of the main PCB. These can be cut using a small sharp knife. The
new links can then be made on the underside of the PCB using tinned
copper wire and a soldering iron − as shown below.
6.4
A310 Issue 1 PCB’s do not have LK12 fitted at all and require more
difficult modification. Please contact Avie for assistance.
6.4
The fitting of RISC-OS 3 will be performed by Avie for £6.50 + VAT −
this charge is simply to cover return carriage.  A
6.4
•   ABC Compiler − I’ve followed the discussions in Archive about the
relative merits of the RiscBasic and ABC compilers with interest
(Archive 6.1 p49). I don’t feel that there is much weight to the
argument that ABC is compiling a language different from Basic V. The
differences are quite small. The evening I received the ABC package, I
got a program of over a thousand lines running with the compiler in
about an hour which included writing some code to initialise several
arrays and removing some redundant code. When I write programs under the
Basic V interpreter, I just take the differences into account.
6.4
My major grumble is that even when I bought the ABC package a year ago,
it had a minor bug which shows itself sometimes when the object code
icon is dragged to a directory − it requires the machine to be reini
tialised. An upgrade to the latest version, which presumably will have
corrected this bug, will cost me £40, which I think is a bit steep if
it’s only the bug fix that is any help to me.  Les May, Rochdale
6.4
•   EasiWord 2 up(?)grade − I did have version 1.07 of Minerva’s EasiWord
and, after seeing the adverts for the WYSIWYG version, EasiWord2, I
decided to go for the upgrade.
6.4
Now, I do not know anything about 1st Word Plus, so the fact that
EasiWord2 was similar conveyed nothing to me. However, when I tried this
new version, I was niggled to find that I couldn’t print without a RISC-
OS printer driver being installed.
6.4
What was much worse was the fact that I could no longer send control
codes to my printer! Horror of horrors − I had lost the ability to print
in double height and in a choice of seven colours on my old Epson
LQ2500+. Let readers be warned!
6.4
I wrote to Minerva and expressed my dismay about their new package. They
very kindly invited me to return version 2 for a refund or part
exchange. I took up their offer and they even gave me a new package of
their old EasiWord 1.07. Well done, Minerva!
6.4
So, before you upgrade, be sure that it actually is an UPgrade for your
purposes.  Keith Lowe, Pickering
6.4
•   Scheme − If Sue Lawley asks me which book I want to take to a desert
island, without hesitation I will have to say ‘Structure and Interpreta
tion of Computer Programs’ by Abelson and Sussman. I reckon that after a
few years of studying it without any distractions, I would know how to
set about programming a computer. The language used by these authors is
Scheme, a stripped down, elegant and refined version of Lisp.
6.4
If you think Lisp is just an acronym for ‘Lots of Irritating Single
Parentheses’, think again. With even a simple parentheses-matching text
editor as supplied with EdScheme, that bogey disappears and you quickly
come to realise that the apparently strange syntax of Lisp-like
languages such as Scheme is in fact a very consistent and clear way of
expressing a computer program. Indeed, just as we in Britain are used to
seeing a Pascal-like pseudo-code used to express a fragment of code, US
books aimed at the academic community often use Lisp for the same
purpose.
6.4
I had read about the first third of the Abelson book before taking the
plunge and buying EdScheme. What arrived was the Scheme optimising
incremental compiler and associated editor on a single disc, a thinnish,
100 pages plus, User Guide and Reference Manual, and a thick, 300 pages
plus, Schemer’s Guide. This package is, in fact, a complete programming
course at a total cost of £60 plus p&p.
6.4
The guide is aimed rather firmly at the late secondary school market and
starts with a deceptively simple approach to programming by asking the
programmer to construct pencil and paper machines capable of performing
given tasks. Even when programming has been introduced, the guide
continues to use the ‘machine’ diagrams alongside the Scheme code.
6.4
This approach has much to commend it. We are familiar with the idea of a
machine which carries out some operation on the raw materials fed into
it. Cans of beans are constructed from a supply of beans and empty cans.
The filled cans are transported as a package but, to make use of them,
we need another machine, a can opener, to open the package and get at
the contents. Much of Scheme programming is concerned with building the
machines to construct packages of data, machines to check their contents
and machines to open up the packages and make use of the contents.
6.4
One quickly comes to realise that within those intimidating parentheses
is the name of a machine which carries out some operation and the names
of items upon which it operates. That applies to arithmetic operations
too, so the label for the operation comes before the operands, so called
Polish or pre-fix notation.
6.4
By chapter 3, you are being asked to write an interpreter to add
fractions (try that in Basic!) and have been introduced to the concepts
of data abstraction and recursion. On the way, you have visited some
slightly eccentric ways of representing numbers in Lisp which, though
really of historical interest only, do force you to sharpen up your
programming skills. Gradually, a mental tool box is built up containing
procedures for selecting, removing, counting, substituting and reversing
elements in lists. The inner workings of the interpreter, ‘The Scheme
Machine’, are explained in chapter 5.
6.4
The machine analogy is explored further in chapter 6 and the problems of
handling infinitely large objects in chapter 7. As a finale, a game is
developed in the last chapter which embraces topics like artificial
intelligence and object-oriented programming.
6.4
Scheme seems to me to be one of the best kept secrets of the computing
world. This implementation is inexpensive and, though not multitasking,
can be entered from the desktop and cleanly exits back to it. It is
hoped that a RISC-OS compliant version will be released during 1993
though the cost will be slightly higher. This will enable Scheme to run
at the same time as other applications and allow code to be dragged into
a Scheme window and compiled. At present, there do not seem to be any
plans for a version which enables the user to create, close or resize
windows from Scheme, though the Macintosh version allows this to be
done.
6.4
I have learned a lot from my experiences with Scheme. Even when
programming in Basic, it has forced me to ask questions like, ‘If I
store data in this way, how much of the code which accesses this data
will I have to rewrite if I decide I want to store my data in a
different way?’ My appreciation of the virtues of a highly consistent
syntax, with few exceptions, has made me more critical of other
languages to which I have access. It has also made me realise just how
much code I have to write which has nothing to do with expressing how I
want the data to be operated on (my model) but is necessary just to tell
the machine how to do it. I think I’m a better programmer for having
taken the time to study Scheme.
6.4
It has also left me with a profound sense of despair. When I was
involved in education, I consoled myself with the notion that, for all
its faults, the A level system produced students who were much better
educated than a High School student in the USA, where the EdScheme
package has its origins. Yet Scheme seems to have made a negligible
impact in this country, probably because those who control education
believe that so long as we can produce GKOs (pronounced Geckos) −
Glorified Keyboard Operators − everything will be all right. With that
in mind, the decision has been taken to ignore the discipline of
programming and concentrate on IT − whatever that is. I think we are
making a big mistake.
6.4
EDScheme is available from Lambda Publications at a cost of £60 plus
£2.50 p&p. A 300 page copyable Resources Pack and a Teacher’s Guide is
also available at additional cost.
6.4
 Les May, Rochdale.  A
6.4

Hints and Tips
6.05
•   A4 pointer loss − Using Impression on an A4, it is very easy to lose
the text pointer. If the pointer is somewhere over the text, it takes
the form of a caret and is very difficult to see. The simple solution is
to click <adjust>. This marks a block of text and your pointer is at one
or other end of the block! Ed.
6.05
•   Ace printer drivers − Many people have asked about the differences
between Risc Developments’ printer drivers and the Ace Computing drivers
(including the Pro-Drivers.) In practice, the standard Ace drivers (£17)
and Risc Developments’ drivers (£23) are the same. However, the Pro-
Drivers (£44) are somewhat enhanced: they support full background
printing, true 24-bit colour, half-toning, dithering and colour
separations. Ace Computing have also recently upgraded their DJ500C
driver to include facilities for the DJ550C. Simon Moy, Archive.
6.05
•   Background printing with RISC-OS printer drivers − Yes it is possible!
RISC-OS printer drivers are capable of printing in the background far
better than third party software due to the fact that the ROM printing
routines are designed to multitask. However, this is not documented and
requires a particular way of printing. Set your printer driver to print
to a file (choose Connection from the Printer Control menu, select File
and enter a pathname to a temporary file). Print the file in question by
the normal method for the application in question − this bit doesn’t
multitask! When it has finished, reset the driver to print to the
parallel (or serial) port. Find the temporary printout file and drag it
on to the printer icon on the iconbar. The printer will begin printing
and you will have full control of the machine. If you are using a RAM
disc or a hard disc, you shouldn’t notice any slowing down at all. Simon
Moy, Archive.
6.05
•   Dating − I can never remember what the time is and so, in common with
a lot of others, I used to have the date and time on the iconbar. This
was not completely satisfactory as it was frequently obscured by
windows. Browsing in the User Guide, I discovered that you could alter
the format of the command line prompt and have the date and or time
displayed before the *. You can do this with:
6.05
SetMacro CLI$Prompt <Sys$Time>, <Sys$Date> *
6.05
This went in my boot file. Now whenever I want the date or time all I do
is, <f12>, read the date and time and press <return>.  Edward Naish,
Gwynedd.
6.05
(On RISC-OS 3, all you do is press <shift-f12> to bring the iconbar
forwards and then <shift-f12> to put it back again. Ed.)
6.05
•   Draw to sprite conversion − In answer to Roger Darlington’s plea in
Archive 6.3, several people gave us the “obvious” way of converting
drawfiles to sprites − you display the drawfile on screen and use the
screen grab facility within Paint to create a sprite. (This emphasises
again the need for a ‘This is Obvious But...’ column, as mentioned in
the Help!!! Column this month. None of the ‘experts’ at N.C.S. thought
of it. Ed.)
6.05
One other solution was mentioned... Drawfile to sprite conversion can be
done with the program Revxtra which is supplied with the Revelation
package. However, this is a copyright program and I don’t know of any PD
equivalent.  Ted Lacey, Southampton.
6.05
To add to that... If you have Revelation ImagePro, you can just drop
drawfiles onto a Revelation picture and save it out again as a sprite −
no need for a separate program.  Gabriel Swords, Norwich.
6.05
•   Easiword − The printed effects (bold, italics, etc) only start working
if you go into !Printers configuration window and, in the Options−Text
Quality, select NLQ rather than No Highlights which is the default.
Steve Hutchinson, Gloucester.
6.05
• Filename dragging − If you have version 1.21 of Edit, you can shift-
drag any object into an edit window and the pathname of the object
appears at the cursor position. This is just the job for constructing
command files. Avoid all those typing errors!
6.05
I read in another magazine that this facility is only available in OS3.
This is not the case, although it does not work with version 1 of Edit.
Edward Naish, Gwynedd.
6.05
•   Fun School − All of the Fun School series (3 and 4) have now been
upgraded to work with RISC-OS 3.10 although some tweaks are still
necessary. For the Fun School 3 range, it is necessary to start in mode
15 before running the programs. In the Fun School 4 range, you will need
to RMKill any VIDC enhancer modules that are running as these cause the
music and graphics to work too fast. Simon Moy.
6.05
•   Inserting RISC-OS (or any other) chips − With the recent introduction
of RISC-OS 3, many Archimedes computers are being returned to dealers
with the pins of the ROM chips damaged. Here is how to avoid that!
Usually, ROM chips are delivered with the pins looking like this...
6.05

6.05
They will not go into the computer’s ROM sockets with the pins splayed
out like this.
6.05
Before handling the ROM chips, make sure that you are not carrying a
static charge. This is easily accomplished by touching an earthed metal
object. You will need a flat area to work on − a formica-topped table is
ideal. Grasp each end of the ROM chip with finger and thumb of both
hands, so that the sharp ends of the ROM pins are facing you horizon
tally. Place the bottom row of pins onto the formica table top and press
down gently but firmly and move the body of the ROM slightly towards
you. Repeat this operation until the pins are exactly at right angles to
the chip body. Turn the ROM over and repeat the above exercise. You
should now have a ROM with the pin layout like this...
6.05

6.05
With the chip just above its socket and the chip notch orientated with
the diagram supplied with your computer, insert one side of the chip
pins slightly into the corresponding side of the socket. Once you are
sure that all pins on that side do align correctly then lower the other
row of pins into the other side of the socket. Again make sure that all
the pins are aligned and then press the chip completely home.
6.05
If you have been unfortunate enough to have broken a pin off of a ROM
chip, there is a solution providing that a small amount of pin is left
on the chip. You will need a “turned pin”, wire-wrap header socket with
the same number of pins as the damaged ROM. Before inserting the damaged
chip into this socket, cut down each pin of the header socket to about
4mm length. Insert the damaged ROM into this socket, and with a very
fine-tipped soldering iron solder a jumper wire from the damaged pin
directly to the empty socket on the header. Once this has been achieved
do not ever remove the ROM from this socket. The completed assembly may
be mounted with care into the original ROM socket. I have repaired
several ROMs by this method.  Mike King, Guernsey.
6.05
•   JPEG − Under RISC-OS 3, JPEG only works providing the hourglass is
switched off. This can be carried out by inserting the SWI call SYS
"Hourglass_Off" at the appropriate place.  Marc Evans.
6.05
•   Paper for inkjets − Daler Superline paper produces excellent results
on my Canon BJ10ex − I bought it a long time ago, so I hope it’s still
available.  S Freeman, Middlewich.
6.05
•   Printer drivers − RISC-OS 3 printer drivers assume that there is
enough “scrap” space on a disc to store any temporary files during a
printing operation. In practice, a hard disc is assumed. If you have a
system without a hard disc, then !Scrap can be either on a floppy disc
or a RAM disc. If there is not enough space for the temporary printout
file, !Printers will crash with no warning. To avoid this, keep !System
and !Scrap on a separate disc on their own. Simon Moy, Archive.
6.05
•   Standard crest in Impression letters − In Archive 6.3 p28, Brian Cowan
asks how he can save floppy disc space with his Impression letters where
he has a college crest, the drawfile of which  takes up 10Kb of memory
per letter − much more than the text − and he fills a floppy disc with
such multiple copies of the crest.
6.05
Surely the answer is to have an empty frame in the standard letter file
into which he drags the drawfile of the crest. All he has to do is to
remember to drag the drawfile whenever he starts a new letter and drop
into the frame a simple dummy drawfile − a single square or circle will
do − that will occupy far less space, before saving to disc. If he needs
to re-load and re-print the document, all he has to do is drop a copy of
the crest in again − surely not a fag. The crest and dummy drawfiles can
be held on the pinboard.  Steve Kirkby, Sutton. (Well, it was his idea
modified by Ed.)  A
6.05

6.05
Using RISC-OS 3
6.05
Hugh Eagle
6.05
I’ve received such a flood of letters in response to my first column
that I probably won’t have time to refer to them all. I apologise in
advance if I don’t use your contribution; I am very grateful for them
all and will try to use them next month. In the meantime, please keep
the hints, tips, comments, suggestions, etc flowing.
6.05
A number of the problems mentioned only seem to affect a few users and
may result from unusual combinations of circumstances which could be
difficult to diagnose. If you think you might be able to help by talking
directly to the people affected, please let me know and I will put you
in contact.
6.05
Fitting the Upgrade
6.05
Several people have commented favourably on the service they have
received from dealers. IFEL has been mentioned a number of times. Nik
Kelly was specially impressed by Simtec’s facilities.
6.05
Götz Kohlberg received his upgrade surprisingly quickly but then had to
wait several weeks for the fitting instructions for his carrier board.
When the instructions arrived, he found the IFEL carrier board very easy
to fit, taking only about 20 minutes. He didn’t have to remove the
backplane or take out the main board.
6.05
Programs That Do Work
6.05
Microdrive – A number of people have noticed that, while the new version
requires you to keep the program disc in the floppy drive while you play
the game, the new courses work with the original program which will run
properly from a hard disc.
6.05
Others – various people have reported that the following work without
any problems:
6.05
FontFX 5.05 (latest upgrade, free from Data Store, recognises the ROM
fonts)
6.05
Flexifile 1.04
6.05
Eureka (has someone found it?)
6.05
Home Accounts
6.05
SigmaSheet
6.05
Arcendium (R. E. Leo says he finds it works, contrary to the report from
Mike Williams in December)
6.05
ArcScan III
6.05
DiscTree 2.30
6.05
Art Nouveau (Tim Nicholson says he uses it a lot, and doesn’t need the
fix I gave two months ago.)
6.05
Investigator 2 works on an A310, but not an A5000.
6.05
Elite (although Raymond Wright says that it did once seize up after
about an hour’s play, which had never happened in RISC­OS 2.)
6.05
Arc-Comm
6.05
Fortran 77 (Acorn)
6.05
Acorn DTP
6.05
Speech!
6.05
Frances Obee is keen to know whether her most used programs will work:
Pendown Plus, Atelier, Prime Art, Chocks Away and Extra Missions, Elite,
Saloon Cars, and Man Utd. Europe. (I can vouch for two on her list: Draw
Plus v2.12 and Lemmings.)
6.05
Richard Burnell says that a small modification will make version 1.00 of
!Define (the Watford VIDC screen mode definer) work: add to line 17760
of !RunImage ELSE tk%=FALSE so that it reads:
6.05
IF INKEY(−256)=162 THEN tk%=TRUE ELSE tk%=FALSE
6.05
Rob Brown says that the original version of Quazer supplied by the
author Julian Rockey, not Impact Software or Fourth Dimension, needs the
screen memory to be set to exactly 160Kb.
6.05
He also says that Corruption and Jinxter will only work with the very
old version of the Shared C Library supplied with the games. You need to
RMKill the RISC­OS 3 version and RMLoad the old version from disc. (He
recommends that you quit any application that uses the Shared C Library
first!)
6.05
He adds that you can run Terramex by first ensuring that “Free” memory
is at least 400Kb then using the command *Run Terracode.
6.05
Programs that May Need to be Upgraded
6.05
From the hints I have received, the general message seems to be that if
a program doesn’t work (and this is more likely to apply to older
programs, especially games) it is worth contacting the supplier or the
manufacturer because they can often supply upgrades (sometimes for quite
nominal sums).
6.05
Among those that have been mentioned are:
6.05
Arcade 3 games compilation from Clares
6.05
Artisan, Artisan 2 and ProArtisan
6.05
Interdictor
6.05
a number of Fourth Dimension games (but Boogie Buggy is apparently
incompatible with an A5000, according to Rob Brown.)
6.05
Droom
6.05
Avon/Murdac (adventure compilation from Topologika)
6.05
ChartDraw, GraphDraw and KeyStrip (excellent PD applications from Chris
Johnson).
6.05
Fonts
6.05
Colin Singleton points out that, although Electronic Font Foundry are
selling (more expensive) “RISC­OS 3” fonts, the old ones work perfectly
well. The only difference is that the new ones have automatic kerning
information built in (for the use of programs that know how to use it).
6.05
Alan Gentle warns that the way FontMax settings are defined has changed.
They used, in RISC­OS 2 to be quoted in pixels, but in RISC­OS 3 are now
quoted in points. The maximum point size for anti-aliasing seems to
refer to the apparent point size on the screen. Thus to make sure that,
say, 15 point text displayed at 130% is anti-aliased he has to set the
appropriate FontMax to 15×1.3, i.e. at least 20.
6.05
David Shepherdson reports that DrawPlus version 2.00 cannot convert
Trinity, Homerton or Corpus to paths and suggests using Draw to do the
conversion and “saving” the result into DrawPlus. (The same seems to be
true of version 2.12. A possible alternative would be to ignore the ROM
fonts and load those you want into a disc !Fonts directory.)
6.05
Marc Evans recommends, contrary to Acorn’s instructions, keeping the
!Fonts directory in a sub-directory, not the root, to save memory and
time if you are not using the disc fonts. (I would suggest you only do
this if you are sufficiently experienced to understand the
implications.)
6.05
John Winter says that Impression (2.17) and LaserDirect (2.05) work as
well – if not faster – with FontSize set to 128Kb.
6.05
(For more on fonts and LaserDirect see the section headed LaserDirect
further on in this column!)
6.05
Filing Systems
6.05
IDE drives
6.05
R. D. Attwood noticed that the hard disc icon on his iconbar had “IDE 4”
underneath it rather than “IDEDisc4”. On trying to access the disc, he
got an error box with “Bad free space map”. Watford Electronics have
advised him that he needs another chip (cost £16) for the hard disc
podule.
6.05
D. T. Software recommended alternative courses of action for one
customer whose IDEFS wouldn’t work with RISC­OS 3.10: either (a)
reformat the hard disc using the formatter supplied with the system or
(b) fit a replacement EPROM incorporating a work-around for a “marginal
infelicity in the way RISC­OS 3 mounts discs.”
6.05
SCSI drives
6.05
C. Purvis has an A3000 with an external 5¼" floppy drive and an Oak SCSI
external interface (fairly old). When he turns his computer on, the SCSI
icon is to the right of the RAM disc’s and when clicked on gives a “Bad
drive” error. He then runs the following file from a floppy:
6.05
*Mount 4
6.05
*RMLoad $.!System.Modules. SCSIFiler
6.05
*RMReinit SCSI
6.05
*RMREINIT SCSIFiler
6.05
*RMReinit SCSIFS
6.05
*ADFS
6.05
after which the SCSI icon disappears from the iconbar. If he then
presses <shift-break> to reset the computer, SCSIDisc4 appears at the
left end of the iconbar and all is well.
6.05
It has been suggested to me that he needs a new ROM on his SCSI
interface. Can anyone confirm this?
6.05
(One thing we have noticed with the Oak SCSI interfaces is that after a
<delete-power-up>, the SCSIdrive configuration is set to 0 and not 4.
What then happens is that the first time you click the hard drive icon,
it says “Bad drive” but then if you click again, it works OK. The
solution is simple: press <f12> and type Configure SCSIDrive 4 <return>
and then do a <ctrl-break>. I think you will find that all is well after
that. Ed.)
6.05
Mike Williams says the Shutdown routine won’t complete unless he has all
his various SCSI drives mounted, so he has to put a disc into his
removable disc drive and spin it up just so that the system can dismount
it for him! He therefore ignores Shutdown, but asks whether there is a
recommended way of safely shutting down a system such as his?
6.05
Mike has also found that if, after switching off, he changes his mind
and switches on again his computer gets stuck half-way through its
initialisation sequence. Is this a RISC­OS 3 feature, a hardware fault,
or just because he didn’t shut down nicely?
6.05
Using “Free space” viewer on SCSI drives
6.05
(This hint from David Lenthall was originally published in Archive 5.5.)
6.05
If choosing the “Free” option from a SCSI drive icon gives an old RISC-
OS 2 style textual report and you would prefer to get the new RISC-OS 3
window with sliders, include the following line in your boot file:
6.05
Set Alias$Free ShowFree -FS scsi %0
6.05
Séan Kelly’s alternative solution (which works beautifully with his A440
and Oak SCSI interface, driver version 1.16) is to copy the SCSIFiler
module from Applications Disc 1 into his !System.Modules directory and
to add the following line to his !Boot file:
6.05
RMLoad SCSI::SCSIDisc4.$.!System .Modules.SCSIFiler
6.05
ArcDFS / DFSReader
6.05
In response to S. J. Furnell’s problem with ArcDFS mentioned in Archive
6.3 p14, Keith Raven recommends !DFSReader on Archive Shareware 31. This
needs one slight adjustment: in line 210 of the !RunImage Basic program,
there is a SYS call that is expected to return 0, 1 or 2 to indicate
whether large or small icons or full directory information are selected
for directory windows. Under RISC­OS 3 this actually returns 64, 65 or
66. If the next three lines of the program are adjusted to reflect this,
it seems to run perfectly. (Richard Burnell suggests adding a new line
instead:
6.05
215 a=a MOD 64
6.05
to achieve the same effect.) (Keith wonders how many similar minor
changes between RISC­OS 2 and RISC­OS 3 will come to light.)
6.05
The Wakefield Acorn Computer Group writes:
6.05
Regarding S. J. Furnell’s problem with ArcDFS there is some good news
here. (These errors are liable to occur when you click on the drive icon
to display a directory window.) The most common error is ‘Bad drive’ and
you can overcome this simply by Naming the disc with a blank name, i.e.
press <space> and then <return>. The ArcDFS Name option is selected via
the ‘Misc’ menu option, followed by ‘Name disc’.
6.05
This doesn’t cure the other (less common) error which is ‘Buffer
overflow’. In that case, the only option seems to be to use the command
line, i.e, press <f12> to get a ‘*’ prompt or bring up a task window
with <ctrl-f12> and type DIR DFS: (not *DISC as you might expect),
followed by DIR :1 to select drive 1, (if required), then CAT for a
directory listing, etc. You can then use the manual *COPY command to
move files around, e.g.
6.05
Copy DFS::1.$.Index ADFS::HardDisc4.$.Junk.MyIndex
6.05
copies a file called $.Index on the DFS drive 1 into the $.Junk
directory of an ADFS hard disc, renaming it MyIndex in the process.
6.05
You can just as easily copy onto SCSI, IDE and RAM discs of course. Note
that you should use *COPY and not *CopyDFS. The very good news is that
operations such as Format, Backup and Verify work fine, as they don’t
need to display a directory!
6.05
5¼" discs
6.05
Peter Prewett has been horrified to find that since upgrading to RISC-
OS 3 he can no longer read any of his 5¼" discs (which he uses for
backing up his hard disc) via his Watford interface. He has tried
resetting the step rate to 6 without success. Luckily, he can still read
discs from his BBC via a DFS to ADFS file transfer.
6.05
Gordon Lindsay-Jones also reports problems with his Watford buffer which
used to work well under RISC­OS 2. Can anyone help? (Have you asked
Watford?)
6.05
Richard Wells initially found that the Beebug DFS Reader refused to work
under RISC­OS 3, but he has now found that if he runs the Commands
utility from the Support disc first then it will run.
6.05
Richard also says he configures his 40-track 5¼" drive (drive 1) with
6.05
*Configure Step 12 1
6.05
and comments that perhaps
6.05
*Configure Step 6 1
6.05
would work for 80-track discs. He says that to restore normal step delay
you should use
6.05
*Configure Step 3
6.05
In response to Roger Power’s query last month, Howard Snow points out
that the syntax is
6.05
*Configure <Step delay> [<drive>]
6.05
Rather confusingly, the information on the step delay given by *Status
is just the step delay for each of the drives, so if drive 0 is
configured to the normal delay of 3, after typing *Configure Step 6 1
*Status would give Step 3 6.
6.05
W. N. Rodger has written with a similar hint and also suggests that
perhaps all floppy drives on a system have to be set if any one is to be
changed.
6.05
Mr Rodger has also noted that he has to leave his Cumana switchable 40/
80 track drive switched to 80-track before selecting Format DOS 360Kb to
format a 40-track MS-DOS disc. When he then runs the PC Emulator, he has
to switch the drive to 40-track to read the disc he has just formatted!
6.05
Compression
6.05
S. R. Freeman reports that, when using Compression version 1.10 with his
A400/1, if he launches an application from CFS and subsequently quits
CFS without first quitting the application, the computer can hang.
Otherwise he has no problems with it.
6.05
Mike Kinghan says that his system hangs when he tries to kill any CFS
filesystem from the iconbar. He only has to do this, however, in order
to free up the Filecore when shutting his machine down. Now he uses the
operating system Shutdown routine (via <shift-ctrl-f12>) instead and has
no problems.
6.05
Mike also observes that “it is not a clever wheeze” to compress CFS
files using the !Squash application supplied with RISC-OS 3. This will
report a gratifying reduction in the logical byte count but when the
actual size of the new file is checked, it will usually be found that
the file has got larger. (Has anyone found a use for !Squash?)
6.05
Dave Livsey has been told by Computer Concepts, in response to a query
about CFS’s use of a temporary file as an intermediate step in the
decompression process, that they “ ... are working on an upgrade to deal
with this ... ”
6.05
Marc Evans says CFS version 1.10 has problems with large jobs – 10
Mbytes, for example – and suggests turning off interactive filer
operations during a large compression, otherwise you may lose files.
6.05
He also writes: Filer_Run can be used in an alias to run CFS’d files
directly from the normal Filer window. To perform this trick, issue the
command:
6.05
Set Alias$@Runtype D96 !Filer_Run CFS#%0
6.05
and just double click on the file to load it.
6.05
Confusing MS-DOS and ADFS discs
6.05
Marc Evans warns that if you regularly use both DOS and ADFS floppy
discs, it is easy to get confused about which is which. You can easily
find yourself copying to a DOS disc thinking it is ADFS; file and
application icons appear in the destination window just as normal and
you can be none the wiser (apart from the copying being rather slower)
until you run out of space after 720Kb.
6.05
Disc corruption
6.05
Keith Raven has experienced two strange floppy disc faults since
upgrading. After deleting the font manager from a PD fonts disc (E
format), he found that three letters from one font had disappeared
although the disc was otherwise intact and verified OK. He was able to
load the font into !FontEd successfully for patching. The other problem
was a “broken directory” on a D format disc, just after deleting a
couple of redundant items (including a !System directory) left over from
RISC­OS 2. The disc verified properly, but a disc sector editor showed
two chunks of continuous Ns in sector 1 (on track 0 head 0). He has
noticed that ADFS buffers is set to 8 (the default?) and wonders whether
there is still a bug there.
6.05
(We reported two months ago Acorn’s telephone help line comment that the
bug which made it advisable to configure the ADFS buffers to zero had
been fixed. We also reported Computer Concepts’ claim that, unless the
buffers are configured to zero, it is vital to dismount disks before
removing them for good from the machine because, otherwise, there may be
trouble later. Does anyone have any proof either way?)
6.05
XOB Remote Logon
6.05
Raymond Wright says that XOB Remote Logon, which allows a BBC Micro or
another Archimedes to access a hard disc attached to an Archimedes via
Econet, is in trouble. He thinks that it is not loading binary files to
the correct locations in the BBC’s memory.
6.05
Printing!!!
6.05
(This seems to be the area that is causing the greatest grief. If anyone
thinks he knows the answers, please get in touch! A number of people
have commented favourably on the Ace drivers. Perhaps, if the people at
Ace think they have got printing sussed, they might be able to help with
some of these problems.)
6.05
(Funnily enough, Tony Cheal of Ace Computing was hoping to write us an
article about printer drivers. As I well know, when you are running your
own business, things can get very hectic. Anyway, if you do get time,
Tony, your fans are waiting in eager anticipation! Ed.)
6.05
Printer buffer
6.05
Several people have commented that the printer buffer doesn’t work
properly. Can this be true or is it that we don’t understand how to use
it properly?
6.05
Richard Burnell says that if he configures the buffer to, say, 300Kb and
tries to print a draw file with lots of text, the computer hangs and not
even the escape key works. (He owns a KX-P1081 dot matrix printer.)
6.05
Mike Kinghan writes: The RISC-OS 3 feature that I had longed for most
was background printing. Running off a manual on my LaserJet II under
RISC-OS 2 could monopolise the computer for half a day. How annoying
then to discover that prints executed from familiar applications are not
automatically run in background under the new OS. A partial remedy,
however, is to use the new Print Manager’s capability to substitute a
file for any installed printer. First take the Printer Control option
from your printer’s iconbar menu. Then select the Connection dialogue
and nominate a file of your choice in place of the installed printer.
This will make applications write to that file when they think they are
printing, and they will do this much more speedily. At your convenience,
redirect the printer connection to your physical printer and drag your
print files to the printer icon. Now they will be printed in background,
allowing you to get on with your work. (But what a palaver! Surely there
must be a simpler way?)
6.05
Change in printer initialisation behaviour
6.05
T. J. Bennett uses First Word Plus version 1 with an A3000 and a Citizen
Swift 24. In order to make best use of this combination, he has written
a !Boot file which redefines screen characters using VDU23 and also
defines some characters not available in the standard printer character
set as download characters to the printer. He has also modified his
printer driver. This all worked in RISC­OS 2, the printer buffer
allowing the booting process to continue to completion whether the
printer was on line or not. In RISC­OS 3, however, the !Boot will not
finish until the printer is connected, switched on and on line. (“A
giant step towards the Neanderthal,” Mr Bennett comments.) He has tried
setting PrinterBufferSize to ridiculous sizes with no discernable
effect, also RMASize and SystemSize. Does anyone have any suggestions?
(I wonder if this problem is related to that described by Mick Day last
month – see Archive 6.4 p9.)
6.05
Mr Bennett has noticed that there is a limit to the amount of data in
the printer driver for defining each character (contrary to what the
manual says) so this cannot be used as the source of download character
data.
6.05
Disc swapping
6.05
Peter Prewett reports that, after loading the !Printer application on
the iconbar and then loading Impression Junior, whenever he tries to
print, he is required to insert the floppy disc from which he loaded the
original printer application. He says he “tried loading system and scrap
from the printer and other discs, none of which makes any difference.”
(My suggestion is that the problem is a misunderstanding of the purpose
of the !Scrap application: the result of running this is simply to tell
the computer that the disc it is on is the place to store temporary
“scrap” files. Thus by “loading” !Scrap from a floppy you tell the
computer to use that floppy for temporary storage during operations like
printing. If you have a hard disc, it is best to copy !Scrap onto its
root directory and to ensure that the first thing you do each time you
turn the computer on or reset it is to open a Filer window on that root
directory so that the location of !Scrap will be “seen.”)
6.05
Peter goes on to comment that system and scrap files are difficult to
use and must be made totally transparent to the end user. “Nobody should
have to set up the computer with these files as it is certainly not
user-friendly.” Does anyone want to volunteer to write a user-friendly
article about system and scrap files? Does anyone have any ideas about
how these might be better organised in a future version of the operating
system?
6.05
Mono greyscales (!?)
6.05
A number of people have noticed that the so-called “mono” option in the
RISC-OS 3 printer drivers actually produces grey scales when rendering
drawfiles (although sprites are printed in black and white as might be
expected).
6.05
LaserDirect
6.05
I may have given the impression in the December column that it is
impossible to print rotated text out of a LaserDirect. John Winter has
pointed out that this can be done fairly easily by creating the text in
Draw and using the Convert to path menu option; this will then print
like any drawfile. Furthermore, the rotated text can be imported into
Impression, again just like any drawfile.
6.05
John Winter has also found that in order to run the RISC­OS 3 !Printers
application after using a Computer Concepts RISC­OS 2 driver, you have
to run the following * commands (either at the command line or in a Task
window or in an Obey or Command file):
6.05
*RMKill PDriver
6.05
*RMReinit PDriver
6.05
Several more people have reported problems with the quality of print
from LaserDirect printers. Michael Lowe says he has considerable
problems printing sprites, mostly incorporated into Impression docu
ments: black areas are often patterned, grey areas sometimes don’t print
and occasionally the whole of the text area breaks up giving a granular
appearance. However, the worst thing, he says, is the unpredictability
of the behaviour.
6.05
There may, however, be an answer to this (but no solution until the new
drivers appear sometime this year, with luck): Computer Concepts have
written to Tim Powys-Lybbe saying: “We are now(!) aware of a problem
with the combination of our LaserDirect drivers 2.05 and RISC­OS 3.1.
The first Landscape page will print correctly, but on subsequent prints,
black output will be speckled. This only occurs when QuickText is off.
6.05
“I would recommend that you turn on QuickText. If necessary, you should
install RISC­OS 2 versions of Corpus and Homerton into your !Fonts
application on disc. If you get a speckled output, changing printing
resolution or reloading the driver will restore a black image.
6.05
“The next release of our drivers should clear up this problem.”
6.05
Tim’s comment is that the output and the problem is a little worse than
their letter concedes.
6.05
(And the problem also occurs fairly predictably on the BJ10 Turbo
Driver. Ed.)
6.05
Alan Gentle recommends inserting a FontRemove statement in the !Boot
file, removing the ROM fonts, and to add Corpus, Homerton and Trinity to
the !Fonts directory on disc, i.e. the boot file should end:
6.05
FontRemove Resources:$.Fonts.
6.05
FontInstall <Obey$Dir>.
6.05
(N.B. the dots at the end of each line are significant.)
6.05
Alan also says Computer Concepts have sent him an update of PrinterLD,
version 2.05s (10-Aug-1992) which seems to work OK!
6.05
When I started printing from Artworks, I found that the greyscales at
600 dpi were rather distinctly graduated. At 300 dpi, the gradation is
much smoother but the individual colours have a much dottier appearance.
On CC’s recommendation, I have found that 600×300 using Screen 2 gives a
good compromise.
6.05
On the question of how long the upgraded printer drivers are taking to
produce, CC said (in a letter dated 17th December 1992): “We have only
just received the source code (we’d have preferred proper documentation,
but there isn’t any) from Acorn. All our drivers require significant
changes to allow full RISC­OS 3 compatibility. Since the RISC­OS 2
drivers work with RISC­OS 3, their conversion is not a priority.”
6.05
Printing after Artworks
6.05
Roger King has found that, after printing from Artworks, computers need
to be switched off (<ctrl-reset> is not enough) before using new
applications that need to print hard copy. If this is not done, spurious
fine horizontal lines of about 4mm length are scattered throughout the
printed page. Has anyone any clues about this?
6.05
Unwanted form feeds
6.05
G. T. Smith’s means of preventing his Brother (Epson FX80 compatible
printer) from spewing out extra pieces of blank paper was to load the
driver into !PrintEdit, delete the formfeed code “12” from the End of
Text Job and to add ESC“8” (“ignore paper empty”) in the Start of Job
line.
6.05
Unwanted line feeds in First Word Plus
6.05
Dave Livsey suggests that the printer driver needs doctoring to remove
the extra line feed that it sends to the printer. Can anyone give
details?
6.05
Editing printer drivers
6.05
G. T. Smith says that he made the mistake of assuming that, having
copied his modified printer driver onto his hard disc, !Printers would
find it and the modification would take effect, but it didn’t. He says
you need to load !Printers onto the iconbar and re-install the driver by
dragging it into the Printer Control window from the !Printers menu.
6.05
Tim Nicholson says that if you change the configuration settings of a
printer which is not the currently selected one (i.e. its icon is greyed
out), then the next time you print, the print manager uses the driver
you have reconfigured not the selected one.
6.05

6.05
PC Emulator
6.05
Mike Small has found that PC programs such as Tetris run first time but
when they get to the point where they ought to re-load, i.e. when he
“dies”, he gets the messages:
6.05
M2ExHandle abort - Bad Entry Sequence
6.05
*CD \Tetris
6.05
FileCore in use (Error number &108A0)
6.05
*Tetris
6.05
He is using version 1.34 of the emulator, patched in accordance with the
Support instructions. (He assumes that the altering of the filetype of
the DOS partition to DOSDisc doesn’t apply to him because he uses
floppies.) He has a user port and MIDI on a basic A3000. Does anyone
have any ideas?
6.05
Nik Kelly finds that his Quick Basic editor scrolls at a useful speed
now that he has an ARM3 and RISC­OS 3, but has found that the freebie,
QB, that comes with MS-DOS 5, does not read his .BAS files.
6.05
He also says that his real-time clock is now always wrong (using PC
emulator 1.6 and MS-DOS 3.3) and asks why?
6.05
Richard Burnell says that version 1.81 of the emulator won’t read Atari
format discs with RISC­OS 3.10 and asks when the new version of the
emulator is due for release.
6.05
Miscellaneous Hardware Tips
6.05
Ground Control teletext adaptor
6.05
Mike Williams found that his problem with this was a hardware problem
and nothing to do with RISC­OS 3. He had changed to an A5000 and there
is a subtle difference in the hardware of the printer port. Ground
Control are offering an upgrade to a podule version − but he doesn’t
know where he is going to find a fifth podule slot!
6.05
SCSI podule and MIDI board conflict
6.05
Richard Burnell found that his Vertical Twist A3000 Econet MIDI/sampler
board would not work with his 8-bit Lingenuity SCSI podule because they
were competing for use of fast interrupts. Version 2.45 of the SCSI
firmware is now available and the MIDI port now works fine. (This
doesn’t sound like a RISC-OS 3 problem to me, but I will include it
anyway!)
6.05
Pineapple digitiser
6.05
The problem Tim Nicholson mentioned in Archive 6.3 p15 was cured by the
addition of an “extra decoupling capacitor”.
6.05
Brainsoft expansion card
6.05
Tim Nicholson reported (Archive 6.3 p15) that this didn’t work. Raymond
Wright says he thinks his works OK, but he has had problems with the
software in RISC­OS 2, some of which have righted themselves in
RISC­OS 3! (The sound module still seems to want to hog all available
memory, though.)
6.05
Maths co-processor
6.05
Raymond Wright says there is no built-in support in RISC­OS 3 for the
maths co-processor card so he still has to load FPEmulator 3.2 from
disc.
6.05
Impression dongle
6.05
Roger Power’s problem mentioned last month turned out to be a defective
dongle (apparently it was one of the very early ones), which CC have
replaced.
6.05
G. T. Smith had a similar problem and while he was waiting for a
replacement dongle it was suggested that he should try “shorting all the
pins of the dongle with a piece of earthed kitchen foil.” Apparently it
worked! (I suggest you only try this if you are confident you know what
you are doing.)
6.05
Roger King reports that he had a problem with both laser printers and a
Deskjet 500 ejecting blank sheets of paper with spurious characters on
the first line of the new page, especially when printing from PipeDream
4 documents. When the Impression dongles were removed the problem went
away.
6.05
Keith Raven agrees that there is something strange. Once, after
switching off the computer and restarting, he tried printing from Basic
using <ctrl-B> but the computer locked up as if the printer wasn’t on
line. Loading !Printers and trying to print from !Edit caused the system
disc to run and the iconbar to flash but no output. Loading Impression
then caused most, but not all, of the previously “printed” output from
!Edit to arrive at the printer. However, he could not repeat this. Since
he had previously used the printer successfully in similar circums
tances, he wonders if the problem might be something to do with printer
initialisation.
6.05
Miscellaneous Hints & Tips
6.05
Screen modes
6.05
Peter Prewett says that mode 31 “flashes fast” on his A440/1 with an
Eizo 9060s Multiscan. He also says that mode 27 will not run on an A310
with an NEC Multiscan. (I also find that mode 31 is unusably flickery on
my Taxan 770 Plus, but surely one of the purposes in giving us such a
wide variety of modes is to give each user a greater chance of finding
some that suit his own particular monitor.)
6.05
Marc Evans recommends non-multisync users to use mode 35 as their
configured mode because it gives a bit more room to work in without
decreasing resolution in the way mode 16 does.
6.05
Desktop Boot files
6.05
Colin Singleton has been puzzled to find that, after creating a Boot
file and rebooting, a number of applications that were running at the
time he created the Boot file failed to start up. The reason for this is
that only “RISC­OS 3 aware” applications will be inserted automatically
into a !Boot file. (Apart from any other reason, the operating system
needs to be able to find out where on disc each application was loaded
from and it can’t do this unless the application is programmed to be
able to tell the operating system when asked. Originally, RISC­OS 3
applications weren’t programmed to behave in this way.)
6.05
If you want your Boot file to run an older application (including, for
instance, current versions of Impression) you have to edit !Boot and add
the appropriate instructions by hand. To do this (assuming you have
already created a Desktop Boot file), load !Boot into !Edit (the quick
way is to <shift-double-click> on it) and you will probably find a
number of lines reading “Filer_Boot ... ” followed by some reading
“Filer_Run ... ”. Immediately after these, add another line reading
“Filer_Run ” followed by the full path name of the application, and
repeat for each application you want to run. The process is described in
more detail in the User Guide starting at page 112.
6.05
Colin Singleton has also tried, as the manual suggests, creating two
!Boot files so that he can switch at will from one “world” to another.
However, when he tries to run them he gets a message “You cannot have
two copies of !Alarm at once.” Does anyone have any suggestions?
6.05
When Götz Kohlberg had created a Boot file and tried rebooting, he got
an error message: “not enough memory or not within desktop world”.
(Could this happen if the configured “language” is not Desktop? (Number
10 in RISC-OS 3.))
6.05
Control of ARM3 cache
6.05
Götz Kohlberg found that the *Cache on command turned the cache off and
*Cache off turned it on! He also found that when he installed the Aleph
One hare and tortoise module, the effect of clicking on the iconbar
icons was also reversed. His work-around for this problem is to include
*Cache on as the very first command in his !Boot file and to include the
command *Cache off immediately before the line that runs the !Arm3
application.
6.05
ROM modules
6.05
R. D. Attwood asks how he can find out which modules are in ROM so that
he doesn’t duplicate them in the system directory on disc. Simple: press
<ctrl-f12> to open a Task window and type RomModules at the star command
prompt. A list of modules in the system ROM and in the ROMS on any
podules installed on the machine will scroll past. To read the list and
print it or save it: bring up the Task window menu and choose the
“Suspend” option then treat the contents of the window just like any
other Edit window.
6.05
To see what else is in ROM ...
6.05
... bring up the menu over the Apps icon on the iconbar and choose the
“Open ‘$’” option. You can now explore all the resources stored in ROM.
6.05
Toolsprites
6.05
Several months ago, the Archive monthly disc included a nice collection
of “tool” sprites designed to smarten up the appearance of window icons
(including things like 3D-effect sliders that look pushed-in when you
drag them). I found that the original set didn’t quite work properly,
but that simply by adding one extra sprite (“blicon22”) copied from a
similar set on a recent Acorn User disc, I got a set which seems to work
fine. This set is on this month’s program disc in a file called
“Win3Tools”. To use it, I suggest you copy it into your !System
directory and include the command
6.05
ToolSprites System:Win3Tools
6.05
in your !Boot file. (Note: this file only includes “22” definitions for
“square-pixel” high resolution modes.)
6.05
Shift key behaviour
6.05
R. D. Attwood has found his left shift key giving a ¤ symbol. On
investigation, he has found that under the key there are two contacts:
one gives ¤, the other £. The !IntKey module on the Support disc seems
to solve the problem but, as he says, it should not be necessary to run
this. Has anyone any ideas?
6.05
Shift-double-clicking into editors
6.05
Marc Evans says that if you shift-double-click on a file to load it into
an editor, the editor can be confused about what filetype it is and can
save an Obey file as Text. (This doesn’t seem to happen with !Edit, but
it does with !StrongEd II. Without the PRM, I can’t be sure but I would
guess that what is happening is that a shift-double-click causes the
Filer to broadcast a message inviting applications to load the file as
if it were a text file. If !Edit picks it up, it presumably then reads
the directory information to find out what the filetype is, whereas
!StrongEd carries on assuming it is a text file. I wonder what other
editors like DeskEdit do?)
6.05
Fix for SciCalc
6.05
Colin Dean, one of the authors of SciCalc, has written in with a fix for
the bug in the +/− operator mentioned in Archive 6.3 p18.
6.05
The +/− operator should always change the sign of the currently
displayed value, unless you are in the middle of entering an exponent,
in which case it should change the sign of the exponent. In practice, it
can produce bizarre results: e.g.
6.05
enter    display
6.05
1E13   1E13
6.05
+/−   1E-13
6.05
=   1E-13
6.05
+/−   3E-30
6.05
The last displayed result should be −1E−13!
6.05
To fix this alter the lines that read:
6.05
5340 IF entry$=“0” THEN entry$= FNtobase(dreg)
6.05
5350 IF base%=1 THEN
6.05
5360 IF (INSTR(entry$,“E”)=0) OR (F%=TRUE) THEN
6.05
to read, instead:
6.05
5340 IF F%=TRUE THEN entry$= FNtobase(dreg)
6.05
5350 IF base%=1 AND F%=FALSE THEN
6.05
5360 IF (INSTR(entry$,“E”)=0)THEN
6.05
Copying a directory into itself
6.05
You can’t any more! (Marc Evans says.) (For those who don’t remember the
heady days when RISC­OS 2 was young, one of its features was that it was
quite easy to fill a disc by recursively copying a directory into
itself!)
6.05
Unplugging !Configure
6.05
!Configure can be unplugged if you do not want the machine altered −
useful in a school situation! (Marc Evans)
6.05
Bug in BASIC64
6.05
A. G. Rimmer reports that there is a fault in BASIC64 in that the format
specification given by @% frequently behaves incorrectly for fixed
format (but not for E or G format). This is illustrated by the following
program:
6.05
10 REM >Double precision test
6.05
20 @%=“F30.2”
6.05
30 PRINT; “With @% = &”;STR$~(@%);“:”‘
6.05
40 REPEAT
6.05
50 INPUT “Number input: ” x
6.05
60 PRINT “ prints as ” x
6.05
70 UNTIL FALSE
6.05
This produces the following output:
6.05
Number input: 102.63
6.05
prints as 1.03E2
6.05
Number input: 1234.5678
6.05
prints as 1.23E3
6.05
Number input: 12.25
6.05
prints as 12.25
6.05
Number input: 0.16
6.05
prints as 1.60E-1
6.05
Acorn have acknowledged the fault and have told Mr Rimmer that there is
no work-around or fix for this problem at the moment.
6.05
Limited precision of SciCalc’s display
6.05
Mr Rimmer also observes that, whereas SciCalc calculates using floating
point numbers with 52 binary digits, which are equivalent to 15.7
decimal digits, you can only input decimal numbers with up to 10 digits
plus a two digit exponent and the program can only display results with
10 significant figures (and 3-digit exponents up to the maximum possible
308). What is wanted is a display for both input and results of a 15 or
16 digit mantissa plus an exponent of up to 3 digits (up to the limit of
308).
6.05
Again Acorn has acknowledged the limitation but has said that a version
with extended display is not available. (I wonder whether Colin Dean
might be able to help?)
6.05
Legible menus in 1st Word Plus
6.05
In Archive 6.3 p12 there was a hint on altering the palette for First
Word Plus release 1.1 so that the menus are legible. This involved a
special Obey file. Gordon Lindsay-Jones suggests a different solution.
He first changed colour 14 to black and saved the altered palette in the
Library directory on the 1wp disc. Before he runs 1wp, he opens the
Library window and double-clicks on this palette, then when he finishes,
he opens the menu on the iconbar palette icon and selects Default to
restore the palette.
6.05
(Mike Williams says version 0.01 of First Word Plus works fine!)
6.05
Using RMFaster
6.05
Mike Williams points out that C programs run much faster if you perform
*RMFaster SharedCLibrary and, similarly, programs that use the FPE a lot
(e.g. ray tracers) will benefit from *RMFaster FPEmulator. Under
RISC­OS 2, these modules were loaded from disc so always ran in RAM. In
RISC­OS 3, we can choose whether to put them in RAM for speed or leave
them in ROM to save memory.
6.05
Draw crashing
6.05
A number of people have had !Draw crash out on them but luckily it saves
work in progress as a valid drawfile in the Scrap directory before it
exits. Therefore, you should be able to recover your work by opening
!Scrap (shift-double-click on its icon), then opening the directory
called ScrapDir inside it.
6.05
Wimp drag and double-click defaults
6.05
Mike Williams finds that settings of 5 for WimpDoubleClickDelay and 10
for WimpDoubleClickMove make detailed work in !Draw much more manage
able. He wonders if anyone knows how to set them exactly like they were
in RISC­OS 2.
6.05
Mouse menu button behaviour
6.05
In response to my tip in Archive 6.3 p19, Tim Nicholson says his mouse
menu button still returns a continuous 2. I then wondered if it was my
tracker ball that was at fault so plugged in my clapped out mouse, but
no: both mouse and tracker ball still return 2 for a moment then 0 when
the menu button is held down. Tim wonders if the mouse response is
configurable.
6.05
Opening a root directory without running all the !Boot files (and
avoiding viruses)
6.05
Tim Nicholson points out that just as control-double-clicking on a
directory icon opens a filer window on the directory without running the
boot files of all the applications in the directory, control-double-
clicking on a drive icon does the same for a root directory. As he says,
this is jolly useful for inspecting discs of unknown origin if you
suspect they might contain viruses.
6.05
Keys that don’t work
6.05
Tim Nicholson says that he “keeps losing his Caps Lock button” and
occasionally <ctrl-shift-f12> doesn’t work. Is this a problem with an
application? (Or has he been dropping his porridge into the keyboard,
again?)
6.05
Task manager
6.05
(Tim Nicholson warns: before trying this save all your work!) Try
clicking <menu> over the Task Manager and choosing the Info option, then
click <menu> in the grey area where it says NAME: VERSION: AUTHOR:.
(Don’t get too excited.)
6.05
Alarm
6.05
Colin Singleton writes, in response to Ken Cowap’s comments on !Alarm
(Archive 6.3 p15): “I think he is wrong in claiming that it contains a
malfunction, but it is certainly less friendly than it might be. It does
indeed appear to insist that the start and finish dates are in the same
year when setting summertime, so you cannot set Off Oct 1992 and On
March 1993. Since we have now finished with BST for 1992, this is not
now a problem.”
6.05
He has found that after setting the two dates for 1993, the Alarm file
contains three dates, the third date being 7th February 1994. He
presumes that at 1 a.m. on that day, an alarm will go off to remind him
to set the summertime dates for that year! Does anyone know what the
date is for?
6.05
Change in behaviour of !Boot files
6.05
Under RISC­OS 2, J. Lageu used to protect directories using the
principle that, whenever a directory was opened, all the boot files in
it were actioned (unless, I think, a sprite with the name of the
application had been loaded into the Wimp sprite pool). He finds that
RISC­OS 3.10 on an A5000 only sees the !Boot file the first time it
“sees” the application. I wonder whether it is possible to tell the
operating system to forget that it has seen the application?
6.05
JPEG
6.05
Steve Dommett has written, in response to Roger Darlington’s query in
Archive 6.3 p28, to point out that there is a very useful implementation
of JPEG included in !ChangeFSI on the Support Disc. The required files
are cjpeg, djpeg and JPEGinfo (which provides instructions). !ChangeFSI
will read JPEG files in the same way as other file types but will need
plenty of spare space on the !Scrap disc. He has converted many sprites,
saving megabytes of disc space, with no appreciable loss of quality.
6.05
Peter Sturdy adds that if you want to use JPEG from the command line, it
is helpful to add the following lines to your !Boot file:
6.05
Set Alias$cjpeg <Obey$Dir>.cjpeg %%*0
6.05
Set Alias$djpeg <Obey$Dir>.djpeg %%*0
6.05
JPEG files can then be manipulated from a Task window without changing
directory or moving the files into the Library directory.
6.05
(The point of JPEG is that it gives much greater savings in filing space
but the compression process is not completely reversible so the
decompressed image will not be identical to the original. For sprites
this doesn’t normally matter.)
6.05
Directory catalogue display
6.05
Raymond Wright has found an irritating change in the display of a
catalogue listing (outside the Desktop): on an 80 column screen you only
get three columns of files instead of 4 and in a 40 column mode you only
get one. Can there be a good reason for the change, he
asks?
6.05
Two Questions
6.05
Backup and the “next” slot
6.05
If you back up an 800Kb floppy with the “Next” slot set to the standard
640Kb, the backup will take two passes. In order to make the computer
copy the whole disc in one go, you have to set the Next slot to more
than 800Kb. Can it be told to use more than the current Next slot if the
memory is available?
6.05
Icon dragging from filer windows
6.05
There is a configuration option (which I can’t find now that I want it!)
that instructs RISC­OS 3 “aware” applications to drag the whole icon
rather than just an empty square box when saving a file. Why on earth
isn’t the Filer programmed to do the same when you drag an icon from a
Filer window?

6.05
Improvements over RISC­OS 3.0
6.05
Paul Skirrow lists the following improvements:
6.05
•   after a Shutdown the Restart button works properly and restarts the
Pinboard correctly
6.05
•   the ADFSBuffers configuration now works correctly and speeds up disc
access significantly (although there may still be a problem in 1Mb
machines).
6.05
•   Paint handles 256 grey scale sprites correctly and also 256 colour
sprites can have any of their colours changed independently (I’d
appreciate an article on the changes in colour handling in RISC­OS 3 –
starting with the basics like the points mentioned here.)
6.05
•   an HP Laserjet / Deskjet bug has been cured.
6.05
•   there is a new X, Y offset facility in the printer driver system,
separated from the “paper margins” (which now specify which area the
printer driver should allow printing in, although it would make more
sense to set the top left values to be the same). Paul says this could
have some other benefits: you should be able to specify negative offsets
to indicate that the printer must move the print head some distance to
actually reach the paper, which might be useful when printing on labels.
Unfortunately, there is a bug: the X, Y origin is not stored and always
defaults to zero! (My mind goes numb when I read this sort of descrip
tion. Can anyone help with an idiots’ guide to printer margins?) (Gerald
Fitton, in Archive 6.4 p39, admitted to not really understanding how it
all worked but gave an empirical approach in his PipeLine column. Ed.)

6.05
•   Printer drivers now set the page length from the “number of text
lines” specification, and a number of other printer problems have been
cured.
6.05
•   Starting up is quicker (10 seconds quicker on an A5000).
6.05
•   Some useful extra screen modes are provided.  A

Hints and Tips
6.6
•   Applications launcher − When working on various tasks on my
Archimedes, I often find that I need to launch more than one
application. For example, I have a wide selection of regularly used
clipart stored in Library format within Draw Plus. These clipart items
form regular input while using Ovation or Impressions. Therefore, when
working on DTP materials I will, on most occasions, be running Draw Plus
as well. The following ‘Pseudo Application’ was therefore devised to
automate the launch of the required cluster of applications (in this
case Ovation & Draw Plus).
6.6
Begin by creating an application directory (i.e. one where the name
begins with a !) with a name to suite the task cluster (e.g. !DTP).
Within this you need to create a !Boot and a !Run file, both Obey files,
using !Edit. The !Boot file only needs to contain the single line ....
6.6
IconSprites <Obey$Dir>.!Sprites
6.6
while the !Run file should contain the following...
6.6
IconSprites <Obey$Dir>.!Sprites
6.6
Desktop -FILE <Obey$Dir>.
6.6
FileLaunch
6.6
In each case, the IconSprites command directs the system to the !Sprites
file for the pseudo-application while the Desktop command in the !Run
file instructs the system to execute a series of commands held within
the file called ‘FileLaunch’.
6.6
Again within !Edit, create a new textfile, to be called ‘FileLaunch’. It
is the contents of this file that will determine which applications are
to be launched by your pseudo-application. For example, the following
will launch both Ovation and Draw Plus as well as opening the directory
viewer to my ‘Files’ folder and my Draw Plus Library folder. The exact
contents being dependent on the organisation of your own system.
6.6
Filer_CloseDir ADFS::4
6.6
ADFS::4.$.MoreApps.!Ovation
6.6
ADFS::4.$.!DrawPlus
6.6
Filer_OpenDir ADFS::4.$.Files
6.6
Filer_OpenDir ADFS::4.$.Files. DrawLibs
6.6
Save this as ‘FileLaunch’ within the pseudo-application.
6.6
The final job is to create an appropriate sprite to represent your new
pseudo-application. This should be 34 × 17 units in size. For mine, I
have merged the small display icons from Ovation and Draw Plus. Once all
is completed, double clicking on the application will, as if it were a
normal application, launch the specified applications and open the
required directory viewers.
6.6
Although this was initially developed on a hard disc based system, it
can be just as useful to floppy based systems. For example, it can be
set up to launch an application along with an appropriate printer driver
or launch a collection of floppy disc based utilities. For users still
working with Risc-OS 2, if the application is given the name !Boot and
placed on the root directory, with the drive set to *OPT 4 2, it gives
many of the same basic facilities as the new Desktop Boot of RISC-OS
3.10.
6.6
NOTE: This system does not effect the normal action when you double
click on an applications file. This will, as always, launch its own
application and load in the file clicked on.  Mike McNamara, Dunstable.
6.6
•   ArcLaser 1.68 and PipeDream − With ArcLaser installed, PipeDream
(3.14) won’t print using its own driver. It is necessary not only to
quit ArcLaser but to RMkill the LQEmulator module.  Richard Torrens,
Cambridgeshire.
6.6
•   BJ10e/ex with CC Turbo Driver − Under Risc-OS 3/3.10 there is a
conflict between the Turbo Driver and the new !Printers system. If you
contact CC, they will tell you to issue the following from the * prompt
....
6.6
RMKill PDriver
6.6
RMREInit PDriver
6.6
A far more elegant way is to place these two commands within the !Run
file of the !Printers application. They should be added immediately
before the line which reads ....
6.6
RMEnsure PDriver 3.16 Error etc
6.6
You will now be able to switch between the two systems without further
problems!
6.6
Mike McNamara, Dunstable.
6.6
•   Dongle problems − Various people have found that, on occasion, they
cannot print from other applications when using an Impression or
Artworks dongle. The situation seems to be worse under RISC-OS 3. The
(temporary) solution is to run Impression or Artworks and quit it. All
should then work OK. (Hint offered by various readers.) (For more
discussion, see the Comment Column on page 18.)
6.6
•   FoxPro version2 − Anyone who has tried to run FoxPro version2 under
version 1.8 of the PC Emulator will have found that it does not work. If
you still have an earlier version of the Emulator, try using that
because it is said to run under some of the earlier versions.  David
Wild, Hemel Hempstead.
6.6
•   Pocket Book power supply − Maplin sell a 300mA regulated transformer
for £8.95 (+£1.20 p&p) which will do the same job as the Pocket Book
mains adaptor and more besides. Maplin’s code is YB23A and the
polarisation should be negative at the tip at 9V.
6.6
You cannot format Flash SSDs above 256Kb unless the Pocket Book is
connected to an adaptor.  Apparently, the Flash SSDs require 12-15V when
formatting.  Mark Goodwin, Worcester.
6.6
(N.B. You can’t just use any old transformer for the PocketBook and we
can’t, of course, be held responsible for any damage which occurs if a
particular power supply does not work. We still cannot get the Acorn
Pocket Book mains adaptors but are buying some Psion ones which we will
be able to sell at the same price − £15 inclusive. Ed)
6.6
• Printers, margins, etc − I disagree with the statement in Archive 6.5
p37 that the X Y origin is not stored. There seems to be a fundamental
lack of understanding in this area. If you want a full explanation, I
will do one but, in brief, the offsets should be set using the technique
described in the readme file in the Printers directory using the TopLeft
printer file. Paper margins should then be set to be at least as big as
the X Y values, then when the printer tries to print, it says to itself,
“left margin 20mm (say) so I have to move the printer head 20mm before I
start printing. Ah, but wait a minute, the print head is already Xmm
from the edge, so I only need to move another (20 − X)mm before I start
printing.”
6.6
The same is true for Y. None of the default definitions I have looked at
are anywhere near right which doesn’t help. BJ330 printers auto sense
the left hand edge of the paper so goodness knows what happens then!
Tim Nicholson, Cranleigh.
6.6
•   Scrap file − Floppy disc users may be pleased to know that the
Wimp$Scrap file, used for in-memory transfer, does not have to be on a
physical medium. I have included the line:
6.6
Set Wimp$Scrap RAM::RamDisc0.$ .ScrapFile
6.6
in my !System.!Boot file. I have configured the machine to automatically
create a RAM disc on switching on and have since avoided much disc
swapping. The !Scrap folder can be avoided, though the Printers (RISC-OS
3) application needs to see it. It can be placed inside the Printers
directory, and the line:
6.6
Obey <Obey$Dir>.!Scrap.!Boot
6.6
included near the beginning of the !Printers.!Run file.  Robert Fuller,
Basingstoke.
6.6
•   !Squash − This program is a side product of the new printer drivers −
the printer definition files are held in squash format. If you change
the filetype of a printer definition file to Squash (filetype_FCA) and
drag it into Squash, it will be unsquashed. Loading into Edit will
reveal a file not unlike the old OS2 PrDataSrc files, although the
graphics data is held differently. RISC-OS 3 requires the definition
files in the unsquashed format. PrintEdit will happily edit files in
either format but always produce Squashed output.
6.6
•   Squirrel and Compression − If, like me, you routinely run your
Squirrel tables from a Compression window, you may eventually get an
error message, “Bad File Format”. If you then open the table’s
directory, you will find that some of the files inside show the white
File_xxx icon.
6.6
Setting the filetype to Squirrel’s DB7 format, will still result in an
error message.
6.6
Here is what Digital Services told me: There is − they think − a bug in
Compression, which causes this effect very intermittently. The thing to
do then is to set the filetype to CFSlzw and then uncompress those
files.  Jochen Konietzko, Köln.
6.6
•   Symbolic links − If you want to have a program in more than one
folder, it is easy enough to store the program once and then create a
link to this program in other folders. In RISC OS 3.1: load !Edit and
create a new Obey file. Shift-drag the program into this window. Save
the Obey file where you want to have the link. Then a double click on
the link-Obey file executes the original program. This works with any
filetype − not just applications.  Thomas Goseberg, Germany.  A

6.6
Using RISC­OS 3.1
6.6
Hugh Eagle
6.6
“Don’t blame it all on RISC-OS 3!” Archive staff have, on a number of
occasions recently, given this advice to Archimedes users. “But it has
only happened since I changed to RISC-OS 3...” Yes, but that doesn’t
mean it’s the fault of RISC-OS 3. Often it’s a hardware fault that was
set off by fiddling with the p.c.b. of an old computer − or, as in some
of the cases in Hugh’s column this month, it is caused by a change in
configuration that occurred when you did a <delete-powerup> (or because
you didn’t do one!). The final reason is that you haven’t read the
manual − if I had a pound for every time I have had to give advice that
was basically, “Read the manual”... Ed.
6.6
Once again I’ve been nearly overwhelmed by the flood of correspondence.
However, do keep the letters coming, please; I’ll try to use them all in
due course, I promise! My address remains: 48 Smithbarn, Horsham,
Sussex, RH13 6DX.
6.6
Paul says, “Try to be brief”, so I’ll get started.
6.6
Programmers’ Reference Manual
6.6
I was told at the BETT show in January that the new PRM will be ready in
“the second quarter” of 1993.
6.6
Programs and hardware that work
6.6
Some additions to the lists of previous months:
6.6
− Chocks Away
6.6
− Conqueror
6.6
− DrawBender
6.6
− “Most 4Mation software”
6.6
− TaskAnsi (from David Pilling) (However, Tord Eriksson adds that you
may want to upgrade it anyway to take advantage of enhancements.)
6.6
− Scanlight Junior 256 (it “seems to thrive”)
6.6
− Pendown and Pendown+
6.6
− PC emulator (“the oldest” version)
6.6
− Pineapple Digitiser (another confirmation that it works, contrary to
the report in Archive 6.3 p15)
6.6
− XOB Remote Logon (Correction!) Ray Wright has found that the problem
he reported last month (Archive 6.5 p29) was in fact a fault in his BBC
Micro, and that Remote Logon works fine. Apologies to XOB.
6.6
− Acorn DTP (Needs to be patched using the Patch from the Support disc.)
6.6
Problem programs
6.6
ShowPage − Tord Eriksson says this “works fine so long as there is no
on-screen text! Not much use for ShowPage therefore, you might say, but
some stunning graphics are possible as shown by the examples supplied
with the program.”
6.6
Atelier − Andrew Campbell has found that Atelier causes his A5000 to
hang whether he uses RISC­OS 3.0 or RISC­OS 3.10. (The sequencer loads
but not the main program. He has tried old modules to no effect.)
However, Atelier does still run on his A3000 after the upgrade to
RISC­OS 3.10.
6.6
The following have been reported as not working:
6.6
Fun School 4 (Under 5’s)
6.6
Satfoot (a satellite tracker from AMSAT-UK)
6.6
Copy Opts (“a handy little utility”) − Tord Eriksson has problems using
his version of ArcFS and the Pinboard: if he glues an arc’d directory to
the Pinboard then tries to open it, the computer freezes! (Perhaps this
is a problem that has been solved in a more recent version of ArcFS.)
6.6
Tord has also found that Ballarena works perfectly but messes up all the
configuration settings (so see the section below called “Saving the
configuration” before you run it!)
6.6
Tord says that he made Powerband work by “editing out lines 430 and
following setting up the voices, and a line further on checking the
voices.”
6.6
Some games, like 4th Dimension’s Saloon Cars, crash with a
“WindowManager:Sprites24” error. What is a “Sprites24” anyway?
6.6
!Psion
6.6
Chris Dawson says that the Psion to Archimedes data transfer program by
Mark Taylor (v 1.65) no longer works with his Psion II. He has tried
everything he can think of, even three different re-wiring
configurations that he has seen. Can anyone help?
6.6
Interrupt (Archive Shareware disc no. 6)
6.6
M. P. Sawle has, up to now, made good use of the Interrupt module, with
the commands *RMLoad Intmodule and *Interrupt 0 in a loading file to
enable subsequent access to the command line with <ctrl-@> from within
both Basic programs and the original version of First Word Plus.
6.6
Having installed RISC­OS 3.10 he finds that:
6.6
•   within Basic programs: <ctrl-@> still works, but <Esc> has the same
effect and this takes priority over the normal programmed effect of the
Escape key at the time.
6.6
•   within First Word Plus: <ctrl-@> works but he cannot return to 1wp
properly (sometimes the current document is lost), also two presses of
<Esc> take him to the command line but the document is always totally
lost.
6.6
•   Lemmings won’t run unless he RMKills the Interrupt module first.
6.6
Can anyone help?
6.6
Investigator II works on the older machines, but not on the A3010/A3020/
A4000/A5000 (presumably because the hardware is different).
6.6
LaserDirect and ROM fonts
6.6
The reason why LaserDirect (and presumably the Turbo Drivers) have
problems with the Corpus and Homerton Oblique fonts in ROM is that the
Outlines files for these fonts do not contain full descriptions of the
letters but simply contain cross-references to the related upright
fonts. (Because these two fonts are simply slanted versions of the
upright fonts, not proper “italic” fonts, all the new font manager needs
to know is where to find the outlines of the upright fonts and what
angle to slant them at.)
6.6
So far as I can tell, it is not necessary to disable the ROM fonts (with
a FontRemove command in the !Boot and/or !Run files of !Fonts) and
include full descriptions of Corpus, Homerton and Trinity on disc (they
would need about 250Kb); I have found that Impression and LaserDirect
manage perfectly well if I just include the following sub-directories in
my !Fonts directory on disc:
6.6
Corpus.Bold.Oblique
6.6
Corpus.Medium.Oblique
6.6
Homerton.Bold.Oblique
6.6
Homerton.Medium.Oblique
6.6
(the files add up to 108924 bytes).
6.6
Turning anti-aliasing off
6.6
Tord Eriksson finds that a, ä and å are rather difficult to distinguish
with anti-aliasing on, so he turns it off by setting FontMax1 to
FontMax5 all to 0, except FontMax 3, which determines the maximum size
of font that is cached and which he sets to 200.
6.6
Replacing the system font
6.6
R. W. Darlington’s prayer in Archive 6.4 p11 has been answered! Rob
Davison has sent in an application (included on this month’s program
disc) called !DeskFonts, which will allow you to replace the system font
on the desktop with an outline font of your choice.
6.6
Electronic Font Foundry RISC­OS 3 fonts
6.6
Colin Singleton thinks that last month’s comment may have been a bit
unfair to EFF in referring to their new fonts as “more expensive”. He
suspects that the old fonts have been reduced in price and the new ones
introduced at the previous price of the old.
6.6
What is “kerning data”?
6.6
Colin Singleton says that he had read about this several times before he
found out what it meant. Kerning is the process of adjusting the space
between letters so that they look more natural. A common example is the
word AWAY. When printed like that, the letters appear too widely spread,
but with kerning the word can be made to look like AWAY. (I have
exaggerated the kerning to make it more obvious.) This can be done
manually in applications like Impression.
6.6
The RISC­OS 3 font manager provides the facility for automatic
adjustment of the spacing between any pair of letters, but in order for
this to work firstly the font has to contain a “kerning table”
indicating the preferred spacing for each possible pair of letters and
secondly the application has to be aware of this facility and to make
the appropriate calls to the font manager. (Archive 5.11 pp5/6 give
details of the font manager calls, showing how you can use them in your
programs.)
6.6
Kerning tables are built into the ROM fonts as well as the newer fonts
from EFF (and other suppliers?) I don’t know which applications use
them; presumably newer versions of Impression, etc. will.
6.6
Filing systems: SCSI
6.6
C. Purvis’s problems (icon in wrong place, “bad drive” error, etc – see
6.5 p26) have been solved by a new ROM for his SCSI podule, supplied
free of charge by Oak.
6.6
Tord Eriksson at first had problems that sounded similar to C. Purvis’s,
then happened to find that his SCSI drive had been “set to 0” and after
“resetting to 4” everything worked perfectly. (I explained this last
month, p27. Ed.)
6.6
Philip Lardner, who has a Brainsoft SCSI podule, is having great
difficulties in copying, moving or deleting files to, from or in the
root directory. It generates all sorts of error messages such as: “Can’t
create ‘SCSI::SCSI.$’ ... ”. Philip has found that he can get round the
problem by copying via the RAM disc and deleting using a trash-can
application but he would obviously be interested in a proper “fix”. (I
don’t know whether it might be relevant but he has put “Set Alias$Free
ShowFree -FS scsi %0”, as recommended last month, in his !Boot file in
order to activate the free space window.)
6.6
Philip has also found that his PC Emulator 1.8 sometimes suddenly hangs
the machine either during or shortly after booting up, or after quitting
the emulator.
6.6
HCCS HardCard 45
6.6
Barry Thompson says this needs a ROM and PAL change by HCCS costing £10
plus VAT.
6.6
IDE
6.6
John Birchenough was initially unable to access his Risc Developments
IDE disc drive. However, with a third (free) replacement ROM fitted in
his podule, it is now working!
6.6
Allan Woods says that the tip given last month for getting the Free
Space window to work with SCSI drives doesn’t work with his Risc
Developments IDE drive. Bernard Perry says he knows someone with an ICS
IDE drive with the same problem. Has anyone got any ideas?
6.6
Barry Thompson says the early Orion IDE drives do not function and
wonders whether the new proprietors at Orion are going to support them.
6.6
Watford 5¼“ drive interface
6.6
Gordon Lindsay-Jones has found that his problem is not with his Watford
buffer, as reported last month. The buffer worked properly with another
disc drive. However, he is still trying to find out what is wrong with
his drive.
6.6
Printing − HP Paintjet, Draw and Poster
6.6
Christopher Price (Sidcup) is having constant problems trying to print
from the RISC­OS 3.10 version of Draw and Poster 1.21 via an Ace printer
driver (he doesn’t know which version) to an HP Paintjet on his A420.
The printer appears not even to acknowledge Draw and refuses to output.
With Poster he gets output sometimes, though white text generally
appears black and the printer “seems to bug out if the artwork involves
clipart” imported from elsewhere. Can anyone help?
6.6
Carriage returns and linefeeds
6.6
Philip Woodward uses a bubblejet printer and often prints direct (i.e.
not via a RISC­OS driver). He finds that if he makes the slightest
error, such as trying to print a non-existent file, when he next tries
to print, everything gets printed on one line. Does anyone have any
suggestions?
6.6
BJ-330
6.6
Tord Eriksson has had numerous difficulties printing on his Canon BJ330.
Using the LQ-860 driver is agonisingly slow (it makes four passes for
each line) and it sometimes gives stripes in graphics areas. He also had
a printer head (costing £200!) written off as a result of a “printout
crashing.” Also he finds that printing can abort before the job is
finished – if he uses 180×180 dpi there are no problems but with 180×360
or higher just part of the page gets printed. He comments that the
printer driver seems to work differently in that it now seems to
calculate the entire bitmap before starting printing and he wonders if
he now needs more than 4 Mbytes just to print an A4 page. Any hints or
tips would be very welcome indeed!
6.6
Tord asks if there is any Archive reader who has tested the plotter
version of the BJ-330.
6.6
Non direct drive laser printers
6.6
Tord also asks for suggestions about laser printers that work well with
RISC­OS 3. Direct drive lasers are out of the question because all his
podule slots are full! What printer drivers exist for Canon laser
printers that use Canon’s own page description language? (How about the
Ace Pro-Driver for LBP4/8 at £44 through Archive?)
6.6
Printing in the background
6.6
Tord Eriksson says that the tip in Archive 6.3 p16 about having two
printer drivers loaded is impossible: if he loads a second !Printers the
first is erased. However, I think he has misunderstood the point: you
should only run !Printers once, but within that application you should
have two drivers loaded and active. (See p62 of the RISC­OS 3 User Guide
for details.)
6.6
Keith Matthews spells out the procedure in more detail: you should drag
two copies of the appropriate printer driver into the Printer control
window of !Printers. Then click <menu> over the Printer control window,
choose the Connection option and connect the first driver to a suitable
file, then give the driver a name ending in ‘f’. The second copy should
be connected to the printer and named with a ‘p’. When the printer
manager is subsequently loaded, two icons appear: the first is selected
by default, so applications will “print” to the file. To produce hard
copy as a background task, just drag the file icon onto the “......p”
icon (you don’t need to select it first.) (I still maintain this is a
ridiculous palaver: why isn’t there a setup option whereby you can tell
the print manager to print first to a file and then automatically send
it to the printer?)
6.6
Keith points out that (although he hasn’t tried queueing printouts) the
printer queue window provides one queue for each printer icon.
Obviously, print image files would have to be created under different
file names and you would need a hard disc to accommodate them all.
6.6
LaserDirect and Turbo Driver news
6.6
Michael Lowe has received a copy of version 2.09 of the LaserDirect
driver which has fixed the speckling problem (see Archive 6.5 pp30/31).
6.6
Computer Concepts have also told both him and Dave Leckie that RISC­OS 3
drivers “are under development and might be available as early as April
but the amount of work involved shouldn’t be underestimated.” Presumably
these will handle rotated text and sprites and work under the RISC­OS 3
printer manager. (Dave Leckie notes, however, that CC have said that
they will not support RISC­OS 3.0.)
6.6
Using a RISC­OS 3 driver after LaserDirect, etc.
6.6
In Archive 6.5 p30 there was a tip that, in order to use a RISC­OS 3
driver after using a RISC­OS 2 driver such as LaserDirect, ArcLaser or
the Turbo drivers you should use the two * commands:
6.6
RMKill PDriver
6.6
RMReinit PDriver
6.6
Maurice Edmundson has pointed out that a convenient place to put these
commands is in the !Run file of !Printers before the first of the
RMEnsure commands.
6.6
Saving paper size settings
6.6
Acorn have advised that there is a fault in RISC­OS 3.0 whereby, if you
alter the page sizes and save them, the new settings are not saved. The
work-around is to <shift-double-click> on !Printers then load the file
PaperRW into !Edit, manually change the settings then save the altered
file. This bug has been fixed in RISC­OS 3.10.
6.6
First Word Plus driver for Deskjet 550C?
6.6
Can anyone supply a printer definition file (Barry Thompson asks)?
6.6
PC Emulator
6.6
John Birchenough’s emulator, version 1.7, will only access the internal
drive A. Even when he makes the default drive B or C, the computer
insists on having a disc in drive A and then operates on the contents of
that drive. Also, it always asks for a disc to be inserted in B even
though it means A. He tried the patch supplied on the Support disc, but
that refused to work with a version higher than 1.6. He has also tried
changing his Config.sys file with variations of the Driver.sys and
Drivparm commands but to no effect. Can anyone help?
6.6
Is RISC­OS 3 faster or slower?
6.6
Some think one, some the other! A number of people have commented that
the machine seems to take longer to redraw the desktop screen or to go
through its initialisation routine. Relying on memory, it is obviously
difficult to tell. Raymond Wright says Pipedream 3.1 appears to take
longer to get spreadsheets loaded. He finds that, initially, he gets a
screen full of zeros which are slowly replaced by real data. (I wonder
if this is because of the way he has got it set up? Robert Macmillan
from Colton said – see Archive 6.3 p12 – that all versions from 3.10 on
work better in RISC­OS 3.) Raymond is sure that RISC­OS 2 was quicker
but he’s not going to reinstall it just to find out.
6.6
Ken Gardner has, however, done just that. He did some controlled timings
before updating to RISC­OS 3.10 and reverted back to RISC­OS 2 to
confirm the anomaly that he found. The timings are as shown below
6.6
For the tests, ADFS buffers were set to zero for both RISC­OS 2 and
RISC­OS 3.10. When Ken changed ADFSBuffers to 2K the time to read the
serial file fell from 48 secs to 19 secs but the Multistore random
access read only dropped from 176 secs. to 170 secs. This is a logical
result but it prompts Ken to ask why we have been instructed to set
ADFSBuffers to zero when this gives so much slower a result than in
RISC­OS 2. Can anyone tell us for sure whether there is a bug or not?
(Is there a danger of losing data if the Buffers are set to a non-zero
value?) (Yes, but only on a 1Mb machine. In 3.11, this has been improved
over 3.10. Ed.)
6.6
Miscellaneous Hints & Tips
6.6
How to avoid continual disc swapping,
6.6
and more on !Scrap files
6.6
Philip Woodward recommends that, on a floppy-disc only system, if you
have several discs for different classes of work and include a !Scrap
file on each, don’t forget to click on the !Scrap icon when you insert a
new disc. If you do forget, you will find that whenever you load a
printer driver, or try to print a drawing, the operating system will ask
you to insert some other disc which has no apparent relevance.
6.6
Draw crashing
6.6
Philip Woodward finds, like a number of others, that Draw crashes quite
frequently reporting an “address exception” error. Does anyone know if
there is anything he can do to avoid this? The error message goes on to
say:
6.6
preserving files in <Wimp$ScrapDir>.Draw before exiting.
6.6
This means that !Draw will attempt to save the work-in-progress as a
file called Draw in the scrap directory. (If it is successful, you will
be able to retrieve the file by opening that directory and double-
clicking on the drawfile icon to run !Draw again and load the file in –
after first making a backup copy of the file if it is important. To find
where the scrap directory is, press <ctrl-f12> to open a Task window and
type the command *Show Wimp$ScrapDir)
6.6
Positioning drawings imported into Draw
6.6
Philip Woodward has also found that, whereas in the RISC­OS 2 version of
Draw, one could superimpose a previous drawing (e.g. one saved on a RAM
disc) in exactly the position it originally occupied in the window by
setting the grid lock on and dragging the file onto the new drawing and
positioning the pointer at the bottom left of the window, this no longer
works. The minimum bounding frame of the old drawing is placed with its
bottom left corner at the mouse position, regardless of the original
position in the old drawing. To reproduce the RISC­OS 2 behaviour, you
have to put something, such as a dummy rectangle, at the bottom left
corner of the old drawing’s window.
6.6
Alarm options
6.6
Philip Woodward has pointed out that, if you include a line in your
!Boot file to set Alarm$Options, you must include % in front of each of
the -format parameters (see Applications Guide p137 for details). Also,
you must put the format string in quotes (which is not, perhaps,
emphasised sufficiently strongly in the Manual.) In both these respects,
it differs from Edit$Options, Draw$Options, etc.
6.6
To illustrate the point, consecutive lines in !Boot might read:
6.6
Set Alarm$Options -format “%w3 %z12:%mi %m3 %zdy”
6.6
Set Edit$Options B12 M99 L1 D
6.6
To get the Alarm$Options to take effect, you also have to choose the
“User defined” option in the Alarm setup window (choose the Setup...
option from the iconbar menu.) It’s not clear from the manual but I
assume that the choice of this option is saved in the CMOS RAM so
doesn’t have to be included in the !Boot file. It’s all a bit confusing,
isn’t it?
6.6
Arm3 hare and tortoise icons
6.6
I didn’t have enough time last month to try out Götz Kohlberg’s tip
(Archive 6.5 p33) about the hare and tortoise module. Now I have, and
I’m totally confused! Whatever combination of commands I use, I seem to
be unable to get the machine to boot up with the cache on without the
hare and tortoise swapping jobs. One thought I’ve had is that I might
swap the names of the hare and tortoise icons inside the !Arm3.!Sprites
file. (Another thought is that perhaps you’re meant to click on the hare
to get the machine to go faster, and that this never worked properly in
RISC­OS 2!) Help!
6.6
Faster filer operations
6.6
If you click <menu> over a filer operation window (e.g. the window that
pops up when you copy files which keeps you informed of progress) and
choose “Faster”, screen update will be less frequent and the filer
operation will be faster. (Marc Evans)
6.6
Copying a large directory to floppy
6.6
If you run out of space when copying a directory or application to a
floppy, insert a new disc, rename it with the same name as the full disc
and click on “Retry”. (Marc Evans) (Someone else has already suggested
that but when I tried it, it said “ambiguous filename”. Ed.)
6.6
<Shift-select> on toggle size icon
6.6
If you click <shift-select> on the toggle size icon, the window will
open as far as it can without obscuring the icon bar. (Marc Evans) (This
seems to depend on the application. Draw windows behave as described,
but Impression windows don’t.)
6.6
False sprites on the Pinboard
6.6
Marc Evans’ Pinboard seems to get confused over icon sizes and sprites
under certain conditions. For instance, it sometimes makes icons three
times bigger and uses the wrong sprites. Tord Eriksson says that in mode
78 (from Computer Concepts NewModes?) some icons disappear, some hide
behind others and some appear twice; a single click on each one restores
peace, but should he have to? Has anyone else found this?
6.6
Configured modes
6.6
Marc Evans has found that *Configure Mode and *Configure Wimpmode seem
to be interlinked – whatever one is configured to the other one follows.
He used to have Mode configured to 0 to save memory when outside the
Desktop, but can’t anymore.
6.6
Tord Eriksson had great problems until he discovered that you can’t
“configure” a non-standard mode. If you want to start up in, say, mode
78 from the Computer Concepts Newmodes module, you have to include in
your !Boot file first a command to load the NewModes module then a
*WimpMode 78 command.
6.6
Icon button types
6.6
A comment for WIMP programmers is that when icons have the double click
flag set, they invert at the first click and group with adjust (like
Filer icons). To get around this, your program must trap for any icon
clicked on and unset the selected bit. (Marc Evans)
6.6
Saving the configuration
6.6
Mike Williams finds that the configuration file saved by !Configure
doesn’t cover all the things he needs to restore after having attempted
to run a rogue program that changes the configuration. He has therefore
built himself a command file containing all the *Configure commands
required to get back to normal plus *RMReinit commands for all the
modules. He created it by doing
6.6
*Spool cfgfile
6.6
*Status
6.6
*ROMModules
6.6
*Spool
6.6
then editing the resulting file into the correct syntax and changing the
filetype to “Command”. (The pukka way of creating such a file is not to
use the pre-RISC-OS *Spool command, but to press <ctrl-f12> to open a
Task window, type Status, <return>, ROMModules and <return>, then press
<menu>, choose the “Unlink” option and save the contents of the Task
window.)
6.6
Several years ago, Risc User published a program called !CMOS_Edit which
allows easy saving and reloading of all the CMOS settings. This seems to
work fine.
6.6
Opening sub-menus automatically
6.6
Richard Hallas has noticed that, if you choose the “Open submenus
automatically” configuration option, then if the pointer is resting to
the left of the “parent” menu item (in the space where a tick would
appear) when the submenu opens it appears on top of the parent menu
rather than to its right. Also, moving around a single menu with several
submenu pointers can have some peculiar effects: sometimes the submenu
pointers are ignored, and it is even possible to have two copies of the
same submenu open.
6.6
Preventing access to !Configure and !Alarm
6.6
Last month we gave the hint that you can prevent users from interfering
with the configuration by unplugging the !Configure module so that they
can’t access the !Configure application. In Acorn’s education newsletter
Arc, there is a hint that you can do the same with !Alarm.
6.6
Opening a directory without booting
6.6
Richard Hallas reports that, if you open a directory with <ctrl> held
down to stop the Filer executing all the !Boot files inside the
applications within the directory, then copying, renaming or deleting
files will cause the directory to be booted unless you remember to press
<ctrl> again.
6.6
Also, if you open a directory with <ctrl> held down and then copy an
application out of it to another directory, then (assuming the Filer has
not already seen the !Sprites file for the application) a “!Sprites not
found” error will be reported. This doesn’t have any ill effects: it’s
just a nuisance.
6.6
Various Matters Arising
6.6
Screen blanker (Archive 6.4 p12)
6.6
Keith Raven has found (in p163 of the User Guide) that *Blanktime sets
the time (in seconds) before the screen goes blank. *Blanktime 0 turns
the screen blanker off. (What I would like to know is: does the
operating system provide any easy means of firing up a graphical routine
as an alternative to the blank screen?)
6.6
Shift key behaviour (Archive 6.5 p34)
6.6
John Woodgate suggests that the £/¤ key may be acting as if it were
pressed as well as the <shift> key. He suggests cleaning with iso propyl
alcohol or proprietary switch cleaner. (For details on how to clean a
keyboard, see 5.1 p25. Ed.)
6.6
Solid sprite dragging (Archive 6.5 p36)
6.6
Barry Thompson writes that the way to turn this on is to use the *
command *FX162,28,3 (It seems really weird to me that a feature like
this, quite a selling point one would have thought, has been implemented
in such a half-hearted way!)
6.6
Spaces in *Set commands
6.6
Brian Fielding has found that whereas in RISC­OS 2 you could use spaces
in *Set or *SetMacro, RISC­OS 3.10 does not allow you to do so:
6.6
SET PipeDream$Path ADFS::Dream_ file.$.,<PipeDream$Dir>.
6.6
was permissible in RISC­OS 2 but you must use
6.6
SET PipeDream$Path ADFS::Dream_ file.$.,<PipeDream$Dir>.
6.6
in RISC­OS 3.10.
6.6
Locked directories in ROM/RAM podule
6.6
Brian Fielding also reports that if you create a New directory in
Computer Concept’s ROM/RAM podule, it creates the directory with a lock.
This lock cannot be removed and the directory can never be deleted
except by re-initialising the Podule.Other problems occur when copying
files into directories on the Podule but these can be ‘SKIPed’. He has
written to CC about this but they have no plans to alter the podule
software to work fully with RISC-OS3.  A

6.6
Timings for various operations (RISC-OS 2 v RISC-OS 3.10)
6.6

6.6
      RISC-OS 2   RISC-OS 3.10   relative
speed
6.6
      (secs)   (secs)   RO3.1 : RO2
6.6
Read a serial file to a string array (50,000 items)   25.4
   48.4   half as fast
6.6
Sort 50,000 strings (using Armsort)   8.0
   8.0   same
6.6
load and sort a Multistore file (2,400 items):
6.6
   load   166   176   slightly slower
6.6
   sort   221   214   slightly faster
6.6
   total   387   390   same
6.6
Print a Draw file (with 32Kb buffer in printer)
6.6
   release computer   500   239
   twice as fast
6.6
   release printer   533   289
   twice as fast
6.6
Mandelbrot (recalculate Clessidra)   90.3
   90.5   same
6.6

Using RISC­OS 3.1
6.7
Hugh Eagle
6.7
I know that RISC­OS 3 has caused some problems, but to claim, as one of
my correspondents does, that the upgrade from RISC­OS 2 to RISC­OS 3 has
caused “more incompatibilities than the upgrade from 8-bit to 32-bit
machines” is taking things just a teeny bit too far. People will be
blaming Acorn for the performance of the England cricket team next!
6.7
I have recently set up an A4000 from scratch. It comes with the
operating system and the RISC­OS 3 Apps already installed, and
everything is beautifully logical and straightforward. If you stick to
up-to-date software and use a standard printer in a straightforward way,
everything is perfectly simple to use. On the strength of this
experience, I have no doubt that, for new machines and new users,
RISC­OS 3 is a considerable step forward. (The A4000 Home Office also
comes with a good wordprocessor, Easiwriter, and with the very easy-to-
use Desktop Database installed and ready to run. It’s a very far cry
from my A310, which was configured to start up on the command line and
it took me two hours to find out how to get to the “desktop”!)
6.7
As for us hardened upgraders, the question whether it was worth the
hassle is harder to answer. Perhaps Acorn should have waited until the
new operating system was more thoroughly tested before they released it,
but then they would have been criticised for keeping us waiting. Perhaps
they should have incorporated more radical improvements, but then the
incompatibility problems would doubtless have been far greater. In this
situation, they couldn’t win!
6.7
Anyhow, rather than complaining, we should direct our thoughts towards
the changes we would like to see next time round. I am accumulating a
number of suggestions for the RISC­OS 4 wish list which I hope to come
back to in a future column; please let me know if you have any to add.
6.7
Enough of the waffle, and on with the business ...
6.7
I am afraid that, again, I haven’t had time or space to use all your
contributions. I will in due course. Many thanks for them all.
6.7
Printing
6.7
Paper X and Y offsets and margins
6.7
This subject seems to have caused a lot of confusion, not least, I
imagine, because there is no reference to the Paper X and Y offsets in
the manual. (These were added after RISC­OS 3.0, I think, and seem to
postdate the manual, although they are mentioned briefly on page 20 of
the RISC­OS 3.10 Release Note.)
6.7
Before I go any further, I feel I should point out that this is a
technical area that the vast majority of users (who are happy to use the
supplied drivers and page settings) never need bother about.
6.7
The following explanation is based on contributions from Bruce Brown and
Tom Hughes and on some of the Read_Me files inside !Printers. It is also
based mainly on experience with HP Laserjets, although most of the
principles should apply equally to most other printers. I hope I have
got it right ...
6.7
The purpose of the Paper X and Y Offsets is to tell the printer driver
which part of the paper the printer is physically capable of printing on
or, in other words, to define where the printer will print if it is told
to print at the top left corner of the paper. These are set via the
!PrintEdit application. The Top and Left Margins, by contrast, define in
which part of the paper the user wants graphics to be printed. These are
set via the Paper Size window in the !Printers application. (It is these
that determine where the grey border appears when you choose the “Show
paper limits” or “Show print borders” option in Draw, Impression, etc.)
6.7
When the printer driver is asked to start printing at the top left
corner of the area defined by the margins: (1) first a printer reset
code is sent, then (2) the start of job codes are sent to the printer,
(3) the printhead (or the imaginary cursor in the case of a laser
printer) is told to move to the top left of the printable area (which,
if the X and Y offsets have been properly defined for the particular
printer, will be the same as the place which they define – i.e. point A
in the diagram below), (4) the printhead will then be moved down by
(Top margin − Y offset) and towards the right by
(Left margin − X offset) to point B, (5) the page will be printed.
6.7
If either of the margins has been incorrectly defined to be less than
the equivalent offset, i.e. so that the formula in (4) gives a negative
result, it will be ignored and the cursor will not be moved at all in
that direction.
6.7
How do you find out what the X and Y offsets should be? If it’s not
clear from your printer manual, the file called TopLeft in the Printers
directory on Applications disc 2 (together with the instructions in the
Read_Me file) might help. Alternatively, you could try deliberately
setting X and Y greater than the margins, so that the cursor movement in
(4) defaults to zero (as explained) and the top left of the image will
be printed at the default printhead position.
6.7
Under RISC­OS 2 only the “margins” could be defined and, for practical
purposes, these had to be set to match the physically printable area.
One of the advantages of the new approach is that if you only want to
print on a small part of the page you can set very wide margins and the
printer driver won’t waste time trying to print nothing (i.e. lots of
white space) in the unused area.
6.7
According to Archive 6.5 p37, Paul Skirrow has suggested setting
negative offsets to force the printer to start printing near the middle
of the page when, for instance, printing labels. Arithmetically this
will achieve the right effect, but not in the intended way! What you are
supposed to do is leave the X and Y offsets unchanged and increase the
margins.
6.7
What will happen if you try to set the margins to less than the paper
offsets? Bruce Brown says you must not do so, because this will cause
the driver to send more data to the printer than it can handle on one
page. The Printers.Read_Me file says that if you do, and if you then try
to print something right up to the top left of the margins, the image on
the paper will be shifted down and to the right of where it is supposed
to be, since you have tried to get the printer to print on the section
of the paper that it physically cannot print on.
6.7
The Printers.Read_Me file also notes that the paper offsets can be
negative. Apparently, the natural print position of some printers is
above the top and/or to the left of the top corner of the paper.
6.7
Contrary to what has been said in this column before, you can apparently
set the X and Y offsets to zero.
6.7
Owen Smith says: “You may be caught out by the way !PrintEdit stores the
paper offsets. Having changed the offsets in the main window, you then
have to open up every graphics resolution one by one and click OK in
each of them and then save the new printer definition file. This is
because the paper offsets are stored in the per graphics resolution data
(in pixels) and the graphics resolution data is encoded when OK is
clicked in the graphics window.”
6.7
Background printing
6.7
It doesn’t work – and that’s official!... I’ve been sent a copy of a
Technical Information news sheet from Acorn, which says: “The printer
buffer module supplied as part of RISC­OS 3 does not work correctly.
Even though the buffer has been configured to a large size, e.g. 2
Mbytes, the buffer module waits for the printer to complete its print
job before returning control to the user. Acorn is currently looking
into this problem and details will be made available if a fix is
produced.”
6.7
Jochen Konietzko, however, wonders what all the fuss is about. He
writes: “On my machine, an A410/1 (ARM 3), with the Ace ProDriver and
the HP DeskJet 500C, there is no problem at all! All I have to do is
configure a sufficiently large printer buffer. I’ve just tried it again
and, with a buffer of 1Mb, I can work in a database with just a slight
reduction in speed.”
6.7
The only reason I don’t use this option is that, unlike the font cache,
it is not possible to drag the task manager slider for the system heap/
stack below the limit set by the configuration of the buffer, so that
the memory is lost completely until a reset.
6.7
My standard setting for the printer buffer is 128 Kb, because that does
not slow printing down (with a 1024 Kb buffer, printing takes about 25%
longer than at 4 Kb, even if I don’t touch the keyboard at all) and yet
I have a kind of “emergency multitasking”; at least, when I decide to
cancel an Impression printout, the button responds almost instantly.
6.7
If I have to print several copies of a text, the print from a file is
better, anyway, because then the printer doesn’t try to hog all
available RAM.”
6.7
Perhaps the reason he has no problems is that the Ace ProDriver doesn’t
use the Acorn buffer module.
6.7
LaserDirect and TurboDrivers
6.7
Computer Concepts say that version 2.09b is now available as a free
upgrade to registered owners, as an interim measure. This fixes a few
minor problems (e.g. the patterning effects and the problems with
sprites with palettes attached) but is not a full RISC­OS 3 version. “CC
have a team working on the printer drivers but the amount of work
involved should not be underestimated – it’s going to take a couple of
months yet.” (This was written on 11th February.)
6.7
Printer driver space requirement
6.7
Acorn advises that you can reduce the amount of disc space taken up by
!Printers by removing (if you have a dot matrix printer) the following
directories from within the !Printers directory: lj, ps and PDumpers.
6.7
Checking if a printer is online
6.7
Acorn also warns of a problem that sometimes affects programs originally
written for the BBC micro. If these use the command ADVAL(-4) to check
if a printer is on- or off-line, this can cause the computer to crash
because the program will not be able to understand the information
returned by the command.
6.7
Problems with VDU2 printing
6.7
Roger Power has used a Basic program for years, which now refuses to
print, causing the computer to hang as if the printer had not been
switched on. The program just has a VDU2 command to cause the screen
output to be sent to the printer and doesn’t use any printer driver. (I
wonder if this problem might be connected to the one mentioned in the
previous paragraph?)
6.7
He then successfully printed a listing from Basic, using <Ctrl-B>, but
when he entered the command:
6.7
VDU2:PRINT TAB(10,5)“Archive” :VDU3
6.7
the word “Archive” was printed on the next line down in the first
column. In other words, the TAB was ignored.
6.7
He has an HP Deskjet 500C. (He normally uses a ProDriver and has no
problems with that.)
6.7
Richard Torrens uses Calligraph’s ArcLaser so doesn’t have a RISC­OS 3
driver installed. He also has a Basic program which uses VDU2 and worked
OK under RISC­OS 2, but now gives an error. He says the solution is to
type
6.7
Unset PrinterType$1
6.7
before entering the program (which doesn’t run in the desktop) and reset
on exit from the program.
6.7
1st Word Plus driver for Deskjet 550C
6.7
Barry Thompson has an answer from Acorn to the question he posed last
month: since the Acorn JP150 printer is an HP Deskjet compatible, use
the 1st Word Plus driver supplied on the JP150 support disc. This driver
also works with Laserjets. This driver supports all the printing effects
from 1st Word Plus but will not support the printing of graphics.
6.7
Programs That Work
6.7
The following programs have been reported as working without any
problems:
6.7
Fun School 4 (Under 5’s) – a new version works with RISC­OS 3
6.7
Manchester United Europe (Peter Young’s son thinks the football plays
faster)
6.7
Intersheet II
6.7
Wordwise A Plus
6.7
Mah Jong, The Game (the latest version – but Peter Young found that it
needed deleting from his hard disc and reloading before it would accept
keyboard input)
6.7
Hard Disc Companion from Risc Developments (but Peter Young has found it
to be much slower)
6.7
Thesaurus from Risc Developments
6.7
Chess, CrossStar and Spell from David Pilling.
6.7
It is interesting to observe that all but two of the programs that
Frances Obee asked about (Archive 6.5 p25) have been reported as working
fine. Of the other two: some problems with Atelier were mentioned last
month but I imagine that the publishers of a leading program such as
this must be able to supply a fix if one is needed; and Prime Art hasn’t
been mentioned at all and I imagine no news is good news.
6.7
Dave Wilcox has sent in a huge list of games that work, which serves to
illustrate two points: first, that most programs do work and second,
that lists of this kind are likely to lead to contradictions since he
reports some programs as working which have previously been reported as
giving trouble.
6.7
Rob Brown has begun compiling a database showing which programs work
(showing version numbers) and reporting compatibility problems that he
is aware of. So far, he has only included programs that he is familiar
with, and already the list is far too long to include in the magazine so
Paul is going to include it on the monthly magazine disc. If you have
anything to add to his database, his address is “Valtanee”, Brighton
Road, Lower Kingswood, Tadworth, Surrey, KT20 6UP.
6.7
Program Problems
6.7
David Holden has offered a couple of tips which may help you to get
programs working, especially demos − which seem to be the biggest
offenders.
6.7
Many demos speed up the system ROMs so that they can run faster. If it’s
a Basic demo, try looking for:
6.7
SYS “Update_MemC”,64,64
6.7
If you find it, either remove it or ‘REM’ it out. This call won’t work
with RISC-OS 3 and will make the computer hang. Symptoms are a complete
lock-up − even the mouse pointer won’t move.
6.7
Another is that the old minimum abbreviation for *CHANNELVOICE was
‘*CHA.’ and for *BASIC ‘*BA.’. These no longer work and require more
letters to be properly identified. There is a utility on one of the
RISC-OS 3 Applications discs which is supposed to cure these problems
but his experience is that this doesn’t always work, so you might need
to actually change the program.
6.7
David adds that RISC-OS 3 is also less tolerant of sloppy disc
identifiers, so filenames should always have a full filing system/path
name.
6.7
He also says that many programs give problems when Acorn’s !Alarm
application is running. This seems to be denied by everyone at Acorn but
it is a fact that he has verified for himself. This applies to a lot of
commercial software as well, so it isn’t just a PD problem but something
that !Alarm is doing. If you normally use !Alarm then try again without
it. If that solves the problem there are plenty of PD equivalents.
6.7
David also points out that many PD programs appear to be written in
machine code because, when you look in the application directory, you
see the ‘Application Code’ icon instead of a ‘Basic Code’ icon. However,
these programs are quite often not written in machine code or a compiled
language but have been disguised by a PD utility and are in fact Basic.
This doesn’t matter unless you want to make one of the changes described
above.
6.7
To restore this type of code to normal Basic, first load it into Edit.
If you have RISC-OS 3 Edit, you just need to hold down SHIFT and double-
click on the program icon to do this, if not you will need to drag the
program icon to the Edit icon. Don’t do anything to the original, but
work on a backup copy in case anything goes wrong. Now look at the first
line of the program. If it is disguised Basic, you will see the words
‘Basic -quit’ somewhere in this line. To change it back to normal Basic,
look for the first [0d] in the file − this will normally be at the end
of the first line. Delete everything up to but not including the [0d] so
that it becomes the first character in the file. Now re-save the file
and change its filetype to Basic (&FFB). If you now try reloading the
program into Edit (assuming you have the RISC-OS 3 version) you should
find that you have a normal Basic program to which you can make the
required changes.
6.7
One cause of compatibility problems that I have found is that !SparkFS
is, by default, configured to use the system sprite area for workspace.
Some older programs expect to have this all to themselves. In these
cases, a simple *SNEW command can help to get the program running.
(Obviously you must first make sure that SparkFS has saved any work-in-
progress!)
6.7
Dave Wilcox writes: “This is probably stating the obvious but, if you
have a hard disc system and normally boot up on the hard disc, it may
help to get games working properly if you configure the machine to Drive
0 instead of the Hard Disc and press <reset>. Also, before starting to
play a game, it is a good practice to open the drive holding the !System
directory, so that it enters the !System path into memory. Also, with
some games, it may be necessary to switch off the cache for ARM 3.”
6.7
Dave Wilcox also says that the following will NOT work under RISC-OS
3.10:
6.7
Powerband Mk2   (4th Dimension)
6.7
Saloon Cars      (4th Dimension)
6.7
Rotor      (Arcana)
6.7
Fireball      (C.I.S.)
6.7
Interdictor I (ver 1.01)   (Clares)
6.7
Corruption      (Magic Scrolls)
6.7
Thundermonk   (Minerva)
6.7
Ibix the Viking   (Minerva)
6.7
Brain Drain      (Minerva)
6.7
Grid Lock      (Minerva)
6.7
Freddy’s Folly   (Minerva)
6.7
Superior Golf & Const. Set   (Superior
Software)
6.7
Filing Systems
6.7
No floppy or hard drive icons?
6.7
Acorn advise: first of all, check that you have the correct number of
drives configured! Secondly, newer machines (A5000 onwards), on startup,
check to ensure that the configured drives are connected and if a drive
is not connected properly its icon may not be displayed − so check the
cable connections to the floppy drive.
6.7
ICS IDE hard drive
6.7
David Shepherdson found that, after he upgraded his A3000, his hard
drive didn’t work. Baildon Electronics, who are the service centre for
ICS, quickly fixed it. He adds, “In fact, as I also have a Calligraph
Laser Podule fitted, this also messed up my hard disc and Baildon fixed
that as well.”
6.7
Watford 5¼“ interface
6.7
Gordon Lindsay-Jones wrote to Watford (as Paul Beverley suggested) about
his problems with their buffer and has had no reply. He has come to the
conclusion that it does not work with RISC­OS 3.10! Can any readers help
by telling him which buffer(s) do work, please?
6.7
SCSI filer
6.7
Seán Kelly has discovered that his tip (Archive 6.5 p27) does not always
work. He writes: “Please allow me to apologise and to explain a proper
way to make the free space window work with SCSIFS.
6.7
“Originally, I advised adding a command to load the new SCSIFiler module
to the desktop boot file. This did work, but only − I have since
discovered − because my desktop boot file was also creating a RAMFS
disc. This forced the initialisation of the replacement module.” He
suggests that the answer is to split the boot process into two stages so
that the SCSI filer is initialised before the desktop is entered. (Tim
Nicholson confirms this.)
6.7
To do this first, rename the existing !Boot file as Deskboot, say, and
create an application directory called !Boot in the root directory and
move the Deskboot file into it. Then create an obey file called !Run
inside !Boot containing the following commands:
6.7
RMLoad SCSI::4.$.!System.
6.7
Modules.SCSIFiler
6.7
Desktop -file <obey$Dir>. Deskboot
6.7
(replacing 4 by the appropriate disc name).
6.7
If you want to smarten up the appearance of the !Boot application, you
can copy the sprites file_fea and small_fea from Resources:$.Wimp.
Sprites into a !Sprites file inside !Boot and rename the sprites !boot
and sm!boot.
6.7
The above is the approach which is also recommended on page 16 of the
RISC­OS 3.10 Release Note. An alternative way of achieving the same end
is (1) to rename the existing !Boot file (the one that contains all the
Filer_Boot and Filer_Run commands, etc) DeskBoot, say, (2) to copy it to
a safe place like the !System directory, (3) to create a new !Boot file
in the root directory, (4) to move into this new file the command to
RMLoad the SCSIFiler module and to add the command:
6.7
*Desktop -file SCSI::4.$.!System.DeskBoot
6.7
(replacing 4 by the appropriate disc name). To my mind, this way is
simpler because it avoids the necessity for creating sprite files,
renaming sprites, etc.
6.7
The $.!Boot.!Run file (under the Kelly/Acorn approach) or the !Boot file
(under mine) can now also be used to load or initialise all sorts of
things before the desktop is started − e.g. extra mode modules, virus
protection, ROM speed up programs, etc.
6.7
Miscellaneous Hints & Tips
6.7
Newlines in !Run files
6.7
P N Cousins says that if there is more than one linefeed at the end of
the !Run file of an application, if he tries “to use the menu button on
the icon” the machine locks up. I presume that he means that, if you
choose the Quit option on the iconbar menu, the application won’t quit.
This is something that I have noticed from time to time. (Invariably,
when it has happened to me, simply pressing <escape> has put things to
rights.) However, I have tried adding newlines galore to the ends of
!Run files without being able to repeat the phenomenon. Can anyone tell
us what is happening?
6.7
On page 16 of the RISC­OS 3.10 Release Note there is a cryptic
instruction that you should not enter any newlines after the second line
in a particular two-line !Run file. I wonder if this is connected.
6.7
Ian Hamilton says that, according to Acorn, Obey files with blank lines
aggravate what he calls a bug which can cause applications not to return
from their Obey files. He suspects that this may be the cause of the
problem that was blamed on Compression in Archive 6.5 p28.
6.7
Newlines at the end of !Boot files
6.7
P N Cousins also says that you must have a newline at the end of a !Boot
file or else the last line will be ignored. So, if you find that an
application which ought to be run by your !Boot file but isn’t, this
might be the answer.
6.7
Hiding the Apps in Resources
6.7
We have mentioned in a previous month that you can hide !Configure and
other ROM-based applications from prying eyes by typing *Unplug
!Configure (or whatever). Acorn advise that, if you do not want users to
see the Resources filing system, you can remove it from the iconbar by:
6.7
*Unplug ResourceFiler
6.7
Setting up an application
6.7
Tim Powys-Lybbe offers the following hints based on how he sets up
applications (using Impression as an example).
6.7
The technique I use is to construct an outer directory to hold (a) the
filer of work for the application and into which outer directory I also
place (b) the application as delivered from the supplier. I copy (c) the
!Sprite file from the application, give the outer directory the same
name as the application and make the !Run file do various things
including opening up the directory of working documents, the application
itself and then back up the working documents on exiting the
application. The !Boot file is more or less copied from the application
though you do need to choose a different system variable name for this
outer directory. (I just add Top to the application’s system variable
name for its directory.)
6.7
I have had several problems getting this working on RISC-OS 3, though
the end result is far better as I now have control over where the
directories open on the desktop and I get a nice clean view of the
Impression start-up picture.
6.7
1. Extension of Filer_OpenDir command
6.7
This command, which opens up a directory on the desktop, otherwise known
to Acorn as a filer, has been extended in RISC-OS 3 to allow you to
decide both where you want to place the filer on the desktop and how
large it is to be (the user guide gives details). For example, in the
!Run file, to open up Impression:
6.7
Filer_OpenDir <ImpressTop$Dir>. Templates 1000 700
6.7
This places the Templates filer with its top left at 1000 OS units
across the desktop from the left and 700 OS units up from the base line.
Acorn define 180 OS units as 1 inch on the screen though this must
presumably depend on the size of the monitor. You will have to
experiment with values for the directory position to get it where you
want on your desktop.
6.7
You can add information about the width and height of the directory by
adding two more numbers:
6.7
Filer_OpenDir <ImpressTop$Dir>. Templates 1000 700 1200 130
6.7
The third number is the width of the directory and the fourth number is
its height. In mode 39, 130 OS units high just accommodates a single row
of the large icon size. (Isn’t the height of an icon, in screen units,
mode-independent?)
6.7
2. Opening an ArcFS directory
6.7
I have had some correspondence with Mark Smith, the ever-helpful author
of this excellent package, and have at last found how to make ArcFS
directories open up from within the !Run file. It is done in two stages,
first:
6.7
OpenArchive <ImpressTop$Dir>. Documents
6.7
Documents is the name of an ArcFS archive that is stored within the
Impression outer directory. Note that ArcFS should already be on the
iconbar for this Open_Archive command to work as it is an ArcFS module
command; you can of course start up ArcFS from within this same !Run
file.
6.7
The second command is:
6.7
Filer_OpenDir ArcFS#Documents:$ 600 1400 1800 260
6.7
The syntax is fairly obvious if you look at the top of the directory for
your archive once it is open!
6.7
3. Backing up to a floppy
6.7
I have a little application tagged on to the end of the Impression !Run
file to cause a back up of the Documents archive after exiting
Impression (or any other application I hasten to add). With RISC-OS 2 it
would work with:
6.7
*COPY <ImpressTop$Dir>.Docu-
6.7
ments :0.Documents
6.7
In RISC-OS 3 this produces an error and one must use instead:
6.7
*COPY <ImpressTop$Dir>.Docu-
6.7
ments ADFS::0.Documents
6.7
Setting up Pinboard without a hard disc
6.7
Brian Fielding has developed an interesting use of the Pinboard, which
he says is especially useful for those without a hard disc:
6.7
After ‘playing’ with RISC-OS 3 for a few days and realising some of the
benefits of the PinBoard, which are obviously more beneficial if you
have a Hard Disc, I worked out a mechanism to fully utilise the PinBoard
so that, when the system was first booted, it would contain all the
applications I would want to access.
6.7
The mechanism described below will display the Applications and when an
icon is double-clicked the appropriate disc will be requested and the
application will be loaded.
6.7
This should be read in conjunction with Chapter 7 of the RISC-OS 3 User
Guide, “Desktop Boot Files”.
6.7
Setting up pseudo-applications
6.7
For each application you wish pinned to or loaded onto your pinboard,
set up a pseudo-application as follows:
6.7
Copy the application, say !AppName, to an empty disc and, using <shift>
and double-click <select>, display the contents of the application.
6.7
Delete all files and directories except !Boot, !Run and !Sprites.
6.7
Edit !Boot to contain the four lines, in the order given:
6.7
IconSprites <Obey$Dir>.!Sprites
6.7
Set AppName$Dir ADFS::DiscName. $.!AppName
6.7
Set Alias$@RunType_fff Run <AppName$Dir>.!Run %%*0
6.7
Set File$Type_fff FileName
6.7
AppName is the name of your application e.g. PipeDream
6.7
DiscName is the name of the disc containing the original application.
6.7
fff is the application’s FileType. (This line should be present in the
original !boot file. There may be other Alias$@RunType_fff lines and
these should also be left.)
6.7
FileName is the Filetype name, e.g. PDream. (This is not essential.
Again there may be others and should be left if desired.)
6.7
Note: <Obey$Dir> must only be used as a pathname to !Sprites. The full
pathname to the original Application must be established, including the
disc name.
6.7
Edit !Run to contain (or create an Obey file):
6.7
Run <AppName$Dir>.!Run
6.7
Do not alter !Sprites. Copy !AppName to your Desktop Boot Disc. Add the
following lines to the !Boot file on your Desktop Boot Disc
‘DskTopBoot’. (See note below if using a RAM disc.):
6.7
Filer_Boot ADFS::DskTopBoot.$. !AppName
6.7
Pin ADFS::DskTopBoot.$.!AppName 100 300
6.7
100 300 represents the position on the PinBoard. With the origin at the
bottom left of your screen, this represents 100 units to the right and
300 units up. Your next application should be placed at 300 300. You
can, of course, pin your applications anywhere on the board.
6.7
If you wish your application to be ready for use, use ‘Run ...’ instead
of ‘Pin ...’.
6.7
RAM disc available
6.7
Unfortunately, the above will require you to load the Boot disc each
time you load any application.
6.7
If you have sufficient space available in memory, create a RAM disc of
at least 64Kb. You need, on average, 7 Kb per application.
6.7
Prepare as above but copy your pseudo-applications to a directory, say,
PseudoApps, on your Boot Disc.
6.7
You will then need to modify your !Boot file as follows:
6.7
Ensure you have a large enough RAM disc.
6.7
Change DynamicArea -RamFsSize 64K.
6.7
Add the line
6.7
COPY ADFS::DskTopBoot.$. PseudoApps.* RAM::RamDisc0. $.* ~CQR~V
6.7
Modify both the Filer_Boot and the Pin (or Run) lines from
‘ADFS::DskTopBoot’ to read ‘RAM::RamDisc0’.
6.7
Notes
6.7
Ensure that you have configured the system to boot from Disc by loading
your Boot disc and executing *Opt4,2 in command line mode.
6.7
Some applications, e.g. Lemmings, are set up to delete RAM disc on
startup and tidy up when QUITed, e.g. UnSet Directories. These should be
altered:
6.7
For Lemmings, the file !Lemmings.!Run should add | at the start of the
line TequeMen and Unset Lem$Dir.
6.7
If memory is a problem, the amount of RAM space can be reduced by adding
a line Delete RAM::RamDisc0.!AppName.!Sprites after the corresponding
Filer_Boot line. This will approximately halve the required space.
6.7
If you have sufficient RAM disc space, it would be helpful to place
!Scrap in it.
6.7
ROM/RAM podule available
6.7
If you have a ROM/RAM podule, with battery backup, then setting up the
pseudo-applications onto the RAM will make application selection much
more user-friendly.
6.7
If you have a ROM/RAM podule, make this your Boot system. Simply follow
the above but change ‘ADFS::DskTopBoot.’ to ‘RFS:’ and remember, in
Command Line mode, to:
6.7
*Configure FileSystem RFS
6.7
*RFS
6.7
*OPT4,2
6.7
*ADFS
6.7
Examples
6.7
The monthly program disc has examples for disc, RAMDisc and for the ROM/
RAM podule.
6.7
Matters Arising
6.7
Backing up in one pass (Archive 6.5 p36)
6.7
Put the following command in your !Boot file:
6.7
Wimpslot -next 800K
6.7
Solid sprite dragging (Archive 6.6 p61)
6.7
(This gets even more bizarre!) Would you believe that, to turn sprite
dragging off, your new friendly operating system requires you to issue
the following commands?
6.7
SYS “OS_Byte”,161,28 TO ,,R2
6.7
R2=R2 AND 253
6.7
SYS “OS_Byte”,162,28,R2
6.7
(Thanks to David Shepherdson for that!)
6.7
!CMOS_Edit: WARNING!
6.7
Bruce Brown warns that, although !CMOS_Edit can be used as I mentioned
last month (Archive 6.6 p60) to save, load and view the CMOS RAM
effectively, if you use the Edit and Update feature, you will be in
trouble. This is because Acorn have now implemented the (previously
reserved) checksum byte at the end of the CMOS RAM. When you “Update”,
!CMOS_Edit will overwrite the correct value with the old incorrect
value. Since the checksum routine is relative, it will then always be
wrong until the power on/delete routine is performed. Until then,
whenever you startup, the operating system will find that the checksum
is wrong and will use default settings rather than those stored in the
CMOS.
6.7
!CMOS_Edit is easily fixed to overcome this: look in the !RunImage
(Basic) file for the loop which writes the values to CMOS RAM and change
the loop counter from 0 TO 239 to 0 TO 238. Leave the loops for saving
and loading alone − they obviously must load and save everything as
previously.  A
6.7
Hints and Tips
6.7
•   Configuring the main Edit window − The RISC-OS 3 version of Edit makes
no provision for preferences concerning the size and the place on the
screen of a newly opened Edit window.
6.7
However, this can easily be rectified. If you run Edit from your hard
disc, the Templates file can be edited according to taste. The nice
thing is that, even from the hard disc, Edit still occupies the same RAM
space (64 Kb on a 4 Mb machine).
6.7
Here is what you do: drag the Edit application from the Apps directory
to your hard disc, click <menu> on the Apps icon, then <select> to open
the root directory, open the Resources directory and copy the three
files inside the Edit subdirectory into your new Edit application. (For
Acorn newcomers, you open an application directory by double clicking
while holding down <shift>.)
6.7
Then drag the Templates file on to the iconbar icon of FormEd (an
application available from N.C.S. on Shareware 20) and drag the Text
window (the one which is filled with a web of diagonal lines) into the
desired position. Save the edited Template file.
6.7
If you always use Edit, don’t forget to change your configuration so
that, after power-on, the correct version of Edit is loaded.
6.7
Jochen Konietzko, Köln, Germany
6.7
•   Faxpack − Did you know that if you manually dial 0336−400−445 and
then, at the prompt, select ‘Receive Fax’, you will receive a 24 Hour
Surface Forecast Chart from The Met. Office. It is very useful if you
want to know what the weather is going to do. Dial 0336−400−401 for the
Marine Index Page of all the forecasts and charts that are available.
Although it is a bit expensive (calls are charged at 36p per minute
cheap rate and 48p per minute at all others times) I find them
invaluable. This only works with FaxPack v2.00 − with earlier versions,
it causes “A Line Error occurred” fault. Steve Monks, Tobermory.  A
6.7

Using RISC OS 3.10
6.8
Hugh Eagle
6.8
Encouragement for the faint-hearted!
6.8
Peter Young says that he is not only electronically illiterate but also
notoriously clumsy and so he left fitting the upgrade for over three
months. When he finally did dare to do it, he found it a lot easier than
he had dreaded − he broke nothing and it all worked first time. The only
worrying bit was changing the links, because the diagram was not clear
but a quick call to the suppliers (Beebug) soon put him right. So his
advice for others is: don’t be scared!
6.8
RISC OS 3.11
6.8
Newer machines and upgrades are now being shipped with version 3.11 of
the operating system. Barry Thompson says that this clears up a NetUtils
problem which occurs when computers equipped with RISC OS 3.10 are used
on a network. The problem was that if you opened or extended a file
longer than 64Kb, anything beyond 64Kb was filled with null characters.
Existing network users can obtain a module (NetUtils) from their local
Acorn dealer.
6.8
Comparing the 3.11 and 3.10 Release Notes, the only other differences I
could find were the following brief notes:
6.8
•    DOS partition size – DOSFS cannot be used with DOS partitions of 32Mb
or larger.
6.8
•    Acorn Lisp – you must issue a *FX 13,11 command before loading Acorn
LISP.
6.8
•    Inspiration – versions of Inspiration up to 1.03 will not run with
RISC OS 3.10 or later.
6.8
(I gather that 3.11 also improves the situation with the ADFSbuffers
problem. On a 1Mb machine with 3.10, you are still advised to set
ADFSbuffers to 0. This is no longer necessary with 3.11. Ed.)
6.8
Programs that work
6.8
Mah Jong, The Game: Peter Young reports that the problem he mentioned
last month was caused by an inappropriate *FX call which he had inserted
in the !Run file – so there seems to be nothing wrong with Mah Jong
itself.
6.8
Hard Disc Companion: Peter Young says that setting an IDEFS cache (see
below) has speeded this up considerably.
6.8
Ovation: In Archive 6.3 p14, we reported a problem with the printing of
indented lines from Ovation. Ken Cowap has now received a new version of
the program (version 1.37S) which, amongst other things, specifically
cures this particular problem. He says the problem turned out to be due
not to Ovation but to the RISC OS 3 printer drivers. Although Acorn
acknowledged this, they saw no immediate possibility of a fix, and Risc
Developments very commendably undertook to rewrite part of Ovation.
6.8
Program problems
6.8
Aldebaran: Atle Mjelde Bårdholt says that you must quit all application
tasks first (enter the Task display and Quit all tasks including
Pinboard), then run the game and it will install itself on the iconbar.
He says another solution is to exit the desktop and then type
*!Aldebaran at the command prompt. (Surely, Aldebaran is one of the most
recent games. Can it be true that it doesn’t run from the desktop?)
6.8
Atle Mjelde Bårdholt also says that old Minerva games generally
reconfigure your machine. To avoid this, skip the !Boot file and click
on the next file in the sequence (e.g. in Hoverbod BS1 and in Missile
Control Title).
6.8
Ralph Gibbons finds that Break147 doesn’t work and would like to know if
there is a fix.
6.8
Mike Cook of Musbury Consultants repeats the point made by David Holden
last month: one reason older programs may not work is that the
abbreviations for some * commands have changed (e.g. *CHA. used to be
*ChannelVoice but is now *ChangeDynamicArea). Running the Obey file
Commands on the RISC OS 3 support disc is supposed to help but if that
doesn’t, he suggests editing the program. This affects the PD sample
player !DSedit and his own !Sample.
6.8
Mike Cook also says that another change in the new operating system
resulted in the sampling part of his software not working. This also
affected the real time sound manipulator !Echo and the !Scope storage
oscilloscope. Any readers with the original programs can send him a disc
for a free upgrade. His address is 5 Helmshore Road, Haslingden,
Rossendale, Lancashire, BB4 4BG.
6.8
Last month, we listed Powerband and Saloon Cars as not working. Rob
Brown says that Fourth Dimension can supply RISC OS 3 compatible
replacements (for £2 per program, he thinks).
6.8
Don’t forget that Rob Brown would welcome any contributions to his
compatibility chart, periodic updates of which will be included on the
Archive monthly program disc. His address is “Valtanee”, Brighton Road,
Lower Kingswood, Tadworth, Surrey, KT20 6UP.
6.8
Printing
6.8
“VDU 2” printing
6.8
David Holden says that, although VDU 2 printing still works with
RISC OS 3, it appears to monopolise the computer until printing is
finished. He has noticed that a couple of his programs that used to
multi-task, no longer do so and BBC emulator programs take a lot longer
to return control to the operator, even if using a printer buffer.
6.8
Printer buffering
6.8
Are the points in the previous paragraph perhaps connected to the non-
functioning of the RISC OS 3 printer buffer mentioned last month? Or are
there further problems? Has anyone got a fix for it or them? The
questions seem endless!
6.8
Ralph Gibbons used to find a printer buffer routine (written by Anton
Calver and on Careware 5 & 7) very useful under RISC OS 2. Does Anton
read Archive? Has he updated his routine to run under RISC OS 3?
6.8
BJ330
6.8
In answer to Tord Eriksson’s plea (Archive 6.6 p56), Brian Cocksedge
recommends the Computer Concepts Turbo Driver (£53 through Archive).
This produces much quicker printouts. The printer should be set to IBM
emulation (he understands this gives a better quality printout). The
stripes in graphics areas are a feature of the BJ printers due to minor
mechanical errors in the feed mechanism and the momentum of the print
head.
6.8
To print A3, as well as setting the paper size in the printer
preferences window, you have to tell the printer the size. Brian does
this by sending “Esc,C,99” to the printer via a small Obey file
containing the commands:
6.8
*FX 138,27
6.8
*FX 138,67
6.8
*FX 138,99
6.8
LaserDirect
6.8
Steve Hutchinson experienced the same spurious lines that Roger King
reported in Archive 6.5 p31. He was told by CC that this was due to the
LaserDirect card not being happy in its podule slot. Rearranging podules
solved it for him.
6.8
Photocopying via LaserDirect
6.8
Tim Powys-Lybbe received version 2.09b of the LaserDirect software,
which cleared up the problems with print quality. However, his scanner
will now not do a photocopy at all, reporting (even with 3 Mb free) that
there is not enough free memory to do a photocopy. If he reverts to
version 2.05a, photocopying continues to work fine.
6.8
Disc drives
6.8
Speeding up an IDE drive
6.8
After he upgraded to RISC OS 3, Peter Young initially found his IDE
drive (with a Beebug interface) to be very slow but eventually
discovered that setting an IDE cache speeded things up considerably. The
command to set it is:
6.8
*Configure IDEFSDirCache 16k
6.8
Before he created the cache, Peter was getting strange messages like dir
“@” not found when he specified directories at the command line with
references like $.xxx... rather than IDEFS:: IDEDisc4.$.xxx.... Since he
is now religiously using full pathnames, he doesn’t know whether
creating the cache might have solved the problem.
6.8
Peter says he still doesn’t really know whether what he wants is a cache
or a buffer. Can anyone throw any light on exactly what they do and what
the optimum sizes are for each in different circumstances? (Since
configuration options are provided, presumably one is expected to use
them, but when and how?)
6.8
SCSI discs
6.8
Andrew Shell encountered the same problems as Philip Lardner (Archive
6.6 p56) with copying, moving or deleting files to, from, or in the root
directory. By resetting his configuration so that *Status reads:
6.8
    SCSIFSDirCache     16k
6.8
    SCSFSdisks         1
6.8
    SCSIFSDrive        4
6.8
    File System    SCSI
6.8
(and using !Configure to set “SCSI hard discs” to 1) he thought he had
solved the problem. All now seems to work correctly so far as saving,
copying and deleting are concerned. However, when he tried the Find
command on the !FONTS directory he got a sequence of finds reading
“...!Boot.!FONTS”, “...!Boot.!FONTS.!FONTS”, ... ,
“.....!FONTS.!FONTS,!FONTS”, etc. He also had a similar problem with the
Stamp command. Does anyone have any solutions?
6.8
SCSIFiler and Set Alias$Free
6.8
Following the hints in past months involving these, P. N. Cousins would
like to know: what is the SCSIFiler module for and how does the Set
Alias$Free hint (Archive 6.5 p27) work?
6.8
Tim Nicholson has noted that using the SCSIFiler module (as recommended
in Archive 6.7 p44 and elsewhere) has the added advantage that it gives
the interactive window for Verify and reads the disc name at start-up
and shows the name under the iconbar icon.
6.8
5¼“ drive interfaces
6.8
Gordon Lindsay-Jones has found, with the help of his dealer, a
replacement for his Watford interface: the one from Risc Developments
works and is, he feels, a superior product. Apart from other advantages,
it has seven switches with which to set the external drive
configurations and comes with instructions on the settings required.
6.8
D. S. Allen and Philip Foster have both written to say that the Watford
interface can be fixed by cutting some wires! D. S. Allen knows of “two
people that have cut the right hand two wires (as viewed facing the
computer front) of the cables from the Watford board to the computer
board and also the one from the Watford board to the internal disc
drive. Both drives work all right after this.” Philip Foster describes
the solution as “cutting through the two end wires (furthest from the
red marker wire and next to the 20 pin removable PAL ROM) of the ribbon
cable running from the mother board to the interface board.” He adds
that “as this is clearly a somewhat irreversible act” he would suggest
ringing Watford to confirm before you do it.
6.8
Miscellaneous hints and tips
6.8
Use of the Copy key
6.8
Richard Torrens writes: “We all know that, in a writable icon, <Copy>
deletes the character to the right of the caret. Did you know that
<Shift-Copy> deletes to the end of the current word and <Ctrl-Copy>
deletes to the end of the current line?”
6.8
How much space for a desktop program?
6.8
P. N. Cousins would like to know how you know how much memory you need
to use when writing programs that run under the desktop. In my
experience, allocating too small a WimpSlot can lead to some most
perplexing error messages, so I would certainly like to know if there is
a foolproof answer. My solution is trial and error: first I estimate how
much space is needed for the program code plus variables and workspace
for sprites, etc, then I set the WimpSlot somewhat larger than my
estimate and repeatedly reduce it (32Kb at a time) until the program
crashes. Not very scientific, but it generally works.
6.8
Miscellaneous
6.8
All these hints are from Rob Davison:
6.8
I have a couple of lines in my application !Run files like this...
6.8
SET Titler$RiscOs3 yep
6.8
RMENSURE UtilityModule 3.10 SET Titler$RiscOs3 nope
6.8
...and near the start of !RunImage, something like the following:
6.8
DIM block% &C00
6.8
SYS “XOS_GSTrans”,“<Titler$RiscOs3>”,block% ,&C00 TO,,l%
6.8
block%?l%=&0D
6.8
IF $block%=“yep” THEN newos%= TRUE ELSE newos%=FALSE
6.8
SYS“XOS_CLI”,“UNSET Titler$RiscOs3”
6.8
•    There are several other configuration options which are not accessible
in !Configure that allow you to setup your system more exactly in the
area of double-click delays, etc.
6.8
One of the more interesting ones which may be useful to those people who
are new to RISC OS, is WimpMenuDragDelay which sets the time in 1/10
second units for which menu activity is disabled after a menu has been
automatically opened. This enables the pointer to move over other menu
entries without cancelling the submenu. To change it, you need to use
the command line (remember that thing?) Syntax:
6.8
*Configure WimpMenuDragDelay <delay>
6.8
•    There is an in-built rolling credit list of all the people involved in
the creation of the Operating System. It is accessed by bringing up the
‘info’ dialog box for the operating system (accessible from the task
manager menu) and clicking with <menu> on the characters ‘t’ ‘e’ ‘A’ and
‘m’ in that order in the string ‘© Acorn Computers Ltd 1992.’ If nothing
happens after a couple of seconds, re-open the menu and try again.
6.8
(Click with <menu>? − that’s against Acorn’s guidelines surely?!)
6.8
However, if you just bring up the dialog box and click <menu> off any of
the icons (e.g. on the grey window background) then the Task Manager
aborts and the only way to recover a normal machine is to press <ctrl-
reset>. (Oops!)
6.8
Boot-up options
6.8
An item on Peter Young’s wish list for RISC OS 4 is the ability to call
different !Boot sequences by pressing user-defined keys at power-on. I
agree that it would be useful to have this built into the operating
system. However, it is already quite easy to achieve a similar result. I
know that some people find Ian Copestake’s !TWO very useful for this
purpose. Alternatively, you can “do it yourself” without too much
difficulty. Assume, for example, that you have saved two Desktop boot
files called Plain and Fancy in a directory called BootFiles. Then
create a Basic file in the root directory called !Boot reading as
follows:
6.8
CLS
6.8
PRINT‘’“Choose one of the
6.8
following sets of
6.8
Boot-up preferences“‘’
6.8
PRINT“   1.  Plain”
6.8
PRINT“   2.  Fancy”
6.8
PRINT‘’“Type 1 or 2”
6.8
A$=GET$
6.8
CASE A$ OF
6.8
 WHEN “1”:OSCLI(“Desktop -File $.BootFiles.Plain”)
6.8
 WHEN “2”:OSCLI(“Desktop -File $.BootFiles.Fancy”)
6.8
 OTHERWISE: OSCLI(“Desktop”)
6.8
ENDCASE
6.8
There are numerous variations on this theme. You can use negative INKEY
values to test for a key being held down rather than waiting for a
keypress with GET$. If you like, you can put Plain and Fancy in, say,
the !System directory instead of creating a new directory, or you can
rename the Basic file !Run, create a directory called !Boot and move
!Run into it along with Plain and Fancy.
6.8
Obviously, you can adapt this idea to suit your purposes, but beware of
pressing R or <Delete> when you switch the computer on, since these will
reset your configuration!
6.8
Splitting the boot sequence into two parts like this was recommended
last month for making sure that the SCSIFiler module is properly
installed before the desktop is initialised.  It also opens up a number
of other possibilities....
6.8
It has always seemed to me that the usefulness of the facility for
automatic saving of desktop boot files is limited (a) because they tend
to get cluttered up with unnecessarily opened directories and booted
applications, and (b) because they omit necessary things like virus
protection modules, extra screen modes, etc. Problem (a) can be reduced
by making sure that you only open those directories that you want open
and boot those directories that you want booted before you save the
desktop boot file (the equivalent of Plain or Fancy in the above
example); (b) can be solved by following the two-stage approach and
putting the more technical bits in a hand-crafted pre-desktop !Boot
file, which you will probably not want to change very often, while the
desktop boot file is left simply to run applications, open directories
and fire up the pinboard.
6.8
So far as I can see, you will still need to do some hand-editing of the
desktop boot file if you want a non“-RISC OS 3 aware” application to be
booted or run automatically. Alternatively, you can get an application
to auto-boot by having the desktop boot file automatically open the
directory that the application is in. Furthermore, you can get the
application to run automatically by editing its own !Boot file to run
the !Run file.
6.8
Tim Nicholson notes that you must also split the boot sequence into two
parts if you want to follow the suggestion in the March Acorn User for
altering the desktop initialisation banner.
6.8
In Archive 6.5 p33, Colin Singleton reported problems with trying to
switch from one “world” to another by double-clicking on a second boot
file. He has concluded that you must reset the computer first, since
running the second boot file doesn’t clear memory first: it simply tries
to add the applications listed in it to those that are already running.
6.8
Finally, to end this section on boot files, a quote from Colin
Singleton, with whom I have a lot of sympathy: “The autoboot facility is
not much use if it does only half the job. I need to be able to
understand what it has produced and to patch it up accordingly. I might
as well carry on using my old boot file, which is what I am doing. The
new feature is useless (and very confusing) for those users who do not
know how to build a boot file. Perhaps in time all software will be
‘aware’ of boot files, then the feature will be useful.”
6.8
ARM3 too fast for Watford scanner (?)
6.8
Mr A. Shell found that, after he upgraded to RISC OS 3 and installed an
ARM3, his Watford Electronics Hand Scanner 1990, Issue 2 (using a Geni
Scan-4500) stopped working. However, he has found that after entering
the command *Cache off, the scanner works perfectly. When the scanning
is finished, *Cache on  speeds the processing up.
6.8
Zooming in Draw
6.8
When I acquired Artworks, I thought one of its neatest features was the
ability to use the mouse to mark the area into which you want to zoom.
Now, I have found while reading the manual (Applications Guide p48) that
Draw does just the same – in fact, it goes one better, because you don’t
need to select a zoom tool first, you just hold down <shift> and drag
with <adjust>. To return to your previous scaling you use <Ctrl-R>; the
manual also describes various other hot-key combinations to adjust the
scaling.
6.8
Turning CapsLock off
6.8
Peter Young has a few old, and some not so old, programs which will only
take upper case input, and putting *FX202,1 into the !Run file meant,
with RISC OS 2, that he didn’t have to remember to press capslock first.
In RISC OS 3, this call seems to turn off keyboard input entirely,
though it occasionally produces input different from what is typed for a
couple of presses before it crashes the machine.
6.8
I think, if I have interpreted the Programmers’ Reference Manual
correctly (pages 511-512 in the RISC OS 2 manual), that the correct code
to turn capslock off while leaving the other keyboard settings unaltered
is *FX 202,16,239; the command *FX 202,16 will turn capslock off but may
alter some of the other settings.
6.8
What *FX 202 does is to alter the “keyboard status byte”. If the command
is given in the form *FX 202,R1,R2 then R2 controls which bits of the
status byte are changed and R1 supplies the new bits. In technical
terms, the new value of the status byte is made equal to
6.8
(the old value AND R2) EOR R1
6.8
Therefore, to force a particular bit in the status byte to adopt value n
(either 0 or 1) you leave that bit unset in R2 and made equal to n in
R1; to leave a bit unchanged you set it in R2 and unset it in R1.
6.8
Capslock is controlled by bit 4 (the one with value 16), so to set it
(thus turning capslock off) you use the values 239 (= 255 minus 16) and
16 for R2 and R1 respectively.
6.8
The shortened form of the command, *FX 202,R1 has the same effect as *FX
202,R1,0
6.8
(It’s not clear to me why *FX 202,1 should have worked either in
RISC­OS 2 or on the BBC.)
6.8
Editing !Run files
6.8
Peter Young asks if one is allowed to amend the !Run files of commercial
programs. My answer is that there is no reason at all why you shouldn’t,
so long as you know what you are doing. You could, for instance, include
a command to turn capslock off (see above) or a command to run a printer
driver automatically. Some programs allow you to change preferences by
setting up system variables in the !Run file. In RISC OS 2, some people
even amended the !Run file of Acorn’s !Edit to make it run on a system
with lots of fonts.
6.8
ROM Speed
6.8
Seán Kelly writes that the old ROM speed up trick still works on some
computers with RISC OS 3.1. On an Archimedes, the ROM speed can be read
with the following Basic lines:
6.8
SYS “OS_UpdateMEMC”,0,0 TO R%
6.8
PRINT (R% AND &C0) DIV &40
6.8
The lowest speed is 0, the highest 2 − the MEMC data sheet defines a
value of 3 as ‘not meaningful’ although it appears to have the same
effect as 2. The speeds are not absolute, but depend on the computer’s
memory clock speed. On old A440, RISC OS 3.1 selects a value of 2, so no
speed up is possible, but on an A5000, a value of 1 is set by RISC OS
3.1 and changing this to 2 increases the machine’s speed by about 5%.
6.8
The ROM speeds can be set by using:
6.8
SYS “OS_UpdateMEMC”,&00,&C0: REM slowest, speed 0
6.8
SYS “OS_UpdateMEMC”,&40,&C0: REM speed 1
6.8
SYS “OS_UpdateMEMC”,&80,&C0: REM fastest, speed 2
6.8
These lines are included as programs on the monthly program disc,
together with the short program called ACK which, on Seán’s un-sped-up
A5000, executes in about 1 second. Double-clicking on one of the speed
changers followed by ACK will show how the different speeds affect your
computer. An A5000 appears to run at speed 2 with no problems. Program
Speed2 can be automatically run by !Boot, providing a slight performance
boost to A5000s.
6.8
RMFaster-ing
6.8
Seán Kelly also reports that *RMFaster on certain modules, such as
Basic, can produce a useful increase in speed for applications that read
the ROM a lot. Without an ARM3, the speed increase can be very large −
with the cache off, he has found Basic programs that run over 30% faster
with Basic in RAM. With the cache on, the increase tends to be about 5%,
so RMFaster-ing on an ARM3 machine is really a waste of memory unless
time is really important.
6.8
Interestingly, both ROM speed-ups and RMFaster-ing only produce a slight
speed increase, typically 5%, on an ARM3 machine. This appears to show
that the cache’s hit rate is very high.
6.8
Also on the program disc ...
6.8
Also on the monthly program disc are a number of utilities from Rob
Davison: SmoothM – a relocatable module which averages mouse coordinates
and is ideal for use in pixel-based art packages such as Paint or
Artisan. To use it, double click on the module and enable averaging with
the command *SmoothON. To restore normal use, enter the command
*SmoothOFF. This module is public domain and is, in fact, based on an
article which appeared in Archive 1.5 (February 1988) by Malcom
Banthorpe.
6.8
SolDrag and NormDrag are short programs to turn solid sprite dragging on
and off.
6.8
Tools and 22Sprites are modified toolsprites and high resolution icon
sprites.
6.8
Some Bugs
6.8
First, from Rob Davison:
6.8
There are quite a number still in the operating system. Most are very
obscure, but some are quite serious and one may be slightly embarrassing
to Acorn! Hopefully, they will do a ‘RISC OS 3 Extras disc’ to fix the
more important ones.
6.8
•    Impression style icons − There is a serious bug in this part of the
window manager. If the application you are writing uses a dialog box
similar to the in-built applications (e.g. Edit) when quitting, but
including the new style icons, the operating system reports an abort on
Data transfer after the application has quit. The solution is to close
and delete all windows yourself and call Wimp_Poll enough times to
ensure that they have all gone before quitting.
6.8
•    Paint − Load the application and repeatedly create 256 colour sprites
of 128x64 pixels. Paint bombs (usually) when the sixth sprite is created
with an ‘Unrecoverable internal error SpriteExtend: Bad pixel
translation table’.
6.8
The only good point is that your file is preserved in Wimp$ScrapDir
before Paint dies.
6.8
The problem mentioned in Archive, where the name in the save dialog box
is corrupted in some circumstances, still exists.
6.8
•    !Edit − This is a fairly worrying one (R. W. Darlington has found
similar problems). Edit has had the same bug since its first release in
RISC OS 2.00. Sometimes, the text window is not redrawn properly and the
caret can appear to be on a different line. Text is deleted or inserted
at the wrong place and it can sometimes be quite a while before you
realise what is going on. I encounter this problem much more now because
of the excellent facilities for editing Basic programs. If Acorn can’t
get windows to redraw correctly in all circumstances, what chance do the
rest of us have? (Quick fix: press <f12>, <return>, <return> before
doing any major editing.)
6.8
•    Operating system bugs − I have one (and only one) 800Kb ‘E’ format
disc which works and verifies fine on RISC OS 2.00 whereas RISC OS 3.00
and 3.10 absolutely refuse to look at it and tell me ‘Disc not
formatted’.
6.8
I have managed to get the dreaded ‘Abort on Data transfer’ while using
interactive file copying but only under extreme provocation. It was much
more common in 3.00!
6.8
There is another problem where, if you set the MOUSE RECTANGLE to allow
the pointer off the left of the screen and then record button clicks,
the y-coordinate is set to −1 with the first returned button click. E.g.
6.8
*POINTER 1
6.8
MOUSE RECTANGLE -1000,-1000,2000 ,2000
6.8
MOUSE TO 100,500
6.8
REPEAT
6.8
MOUSE x%,y%,b%
6.8
PRINTTAB(0,0);“x=”;x%;“  ”
6.8
PRINTTAB(0,1);“y=”;y%;“  ”
6.8
PRINTTAB(0,2);“b=”;b%;“  ”
6.8
UNTIL b%<>0
6.8
END
6.8
Run this and move the pointer off the left of the screen. Click a button
and the displayed y coordinate changes to −1 regardless of the last
position. (The solution is to do another mouse read immediately after
the UNTIL and use the values returned from that.)
6.8
Next, from Seán Kelly:
6.8
!Paint has two bugs related to deleting rows and columns in sprites. The
first is that when columns to delete are selected by dragging right with
the mouse, it is not possible to delete the rightmost column of the
sprite (no matter how far it is zoomed in). The second is that, when
dragging rows to delete downwards, the topmost row (where the operation
was started) is not deleted.
6.8
If there is data in the RAM filing system, RISC OS 3.1 complains if you
attempt to leave the desktop by selecting EXIT from the task manager
menu, although the RAMFS is not affected.
6.8
Sometimes, a RAMFS filer window stays on screen when the RAMFS has been
killed.
6.8
Finally, from R. W. Darlington:
6.8
When I load so many items on the iconbar that it has to scroll to reach
opposite ends and I then delete an item from the middle of the list, I
sometimes finds that as well as the deleted one disappearing, other
icons disappear leaving only their names. Also, all the directory icons
and application icons disappear from directory viewers leaving only
their names as reminders as to what is there.
6.8
In !Paint, when I try to x-scale or y-scale a sprite, I find that it
takes one column off the right hand edge and puts it at the left hand
edge.
6.8
Has anyone else experienced these or other bugs?
6.8
Matters Arising
6.8
Alarm (switching between GMT and BST)
6.8
(Archive 6.3p15 / 6.5p35) Colin Singleton confirms that the alarm set
for 1 a.m. on 7th February 1994 is to remind you to set the BST on and
off dates in 1994.
6.8
Don’t worry: you don’t have to stay up to catch this alarm: it will come
up when you next switch the computer on.
6.8
Ken Cowap has sent in a description which he hopes will complement the
rather brief instructions in the Applications Guide (p132) on this
rather convoluted facility:
6.8
When this option is selected, Alarm will automatically switch between
GMT and BST times on the dates you specify. Click on the arrows to
change the time, day, month and year.
6.8
Entries must be for one calendar year only even though the “BST starts
on:” date may already have passed. Any such past date will be ignored
but must be entered nevertheless.
6.8
As the changeover dates are liable to undefined variation from year to
year, Alarm will issue a warning during the month prior to that of the
next year’s first expected changeover.
6.8
The warning will be recorded in the Alarm Browser Window as “Application
alarm set by Alarm”, and the dates/times of valid changeover(s) will
still be activated. These alarms appear in dark grey and cannot be
selected. When the warning is activated, a message appears saying: “The
dates for entering and leaving BST are no longer valid and should be
reset as soon as possible.” and two options are offered: “Reset them
later” and “Reset them now.” The warning will be repeated automatically
at 24 hour intervals until “Reset them now” is selected and the dates
and times are reset.
6.8
Ken warns that anyone wanting to experiment by setting the system clock
into the future should take care to back up their alarms first, since a
“move into the future” will activate and lose all alarms set to go off
before that date. Also note that there is a long delay accompanied by
unexplained hourglass activity before the warning notice appears.
6.8
Impression dongle
6.8
(Archive 6.5 p32) Colin Singleton has been assured by Computer Concepts
that shorting the pins of the Impression dongle can do no harm.
Switching the computer off while Impression is running can cause
problems. Apparently, the reason is that Impression continually checks
that its dongle is present and, if it happens to be checking when power
is lost, the dongle is liable to receive a burst of static which renders
it inoperable. This will drain away if you leave it for a time (perhaps
a few hours) but the quick cure, according to CC, is to short the lines
to each other using a small coin rubbed against the rows of male pins.
Colin recommends removing the dongle from both the machine and the
printer first.
6.8
Getting the hare and tortoise into step
6.8
(Archive 6.6 p59) Jim Nottingham has written in response to my plea...
“the solution is to fire up the computer in slow mode with a *Configure
Cache off command and run the !Arm3 application in the Boot file. This
brings up the computer with the cache on and the hare on the iconbar. If
you want the computer to boot up with the Arm3 cache off, add *Cache off
to his Boot file after running the Arm3 application.”
6.8
Well! I must be completely dense. I’ve tried to follow these
instructions to the letter. I have typed *Configure Cache off at the
command line (and *Status confirms that the cache configuration is
“off”) and I have a !Boot file (file type “Desktop”) the first two lines
of which read:
6.8
Filer_Boot ADFS::Four.$.!System
6.8
RUN system:!Arm3
6.8
When I boot up, the cache is off and the tortoise on the iconbar. When I
click once on the tortoise, the cache is turned on, but the tortoise
stays. When I click a second time, the hare appears but the cache is
turned off. Adding Cache on as the next line of the !Boot file doesn’t
seem to help. What do I do next (apart from swapping the names of the
two icons as I’ve suggested before!)?
6.8
Pinboard backdrop sprites
6.8
In Archive 6.4 p12, R. W. Darlington gave a suggested modification to
the !Boot file which results in a different backdrop being shown each
time the computer is reset. His method causes the available sprites to
be chosen in strict rotation. Geoffrey Rimmer has suggested a variation
of the same idea which would choose one of the sprites at random.
6.8
Assuming, as before, that there are 75 sprites called Sprit00 to Sprit74
in the directory PinSp, his suggested code is:
6.8
Set S$ <Sys$Time>
6.8
SetEval S$ S$ RIGHT5
6.8
SetEval NN STR(VAL(S$ RIGHT2 + S$ LEFT2)MOD 75)
6.8
If LEN(NN)=1 Then SetEval NN “0” +NN
6.8
SetEval Tile$Name “Sprit”+NN
6.8
Pinboard
6.8
Backdrop -T ADFS::4.$.PinSp. <Tile$Name>
6.8
Finally ...
6.8
Page 140 of the User Guide says that the example commands to set
Alias$@PrintType and Alias$@RunType variables must have trailing spaces.
Does anyone know why?  A
6.8
Supermarket − In Four Languages
6.8
Alex Thomas
6.8
This review is a language teacher’s perspective on the ‘Supermarket’
language programs produced by New Era Software in four languages
(English, French, German and Spanish).
6.8
Surely, I thought, this must be every modern language teacher’s dream
software for the Archimedes. Just think of the realistic situations set
against the background of shopping in a supermarket which could meet the
demands of the National Curriculum (Modern Languages and IT) and the
GCSE syllabuses. This disc must, I thought, deal with asking for
particular items, quantities, containers, problems at the checkout,
queries about the bill, breakages, complaints, special offers, coupons,
shoplifting, advertising, job vacancies − the potential is tremendous.
6.8
With my mind buzzing with such expectations, I was eager to try out this
program with a class, so I sat down after school to run through it and
plan some meaningful tasks. I was hoping to use the German version, so
that was where I started. I needed to enter the shop name and a message
to go on the till receipts − everything was in German, so I was feeling
very positive about the program. (The function keys are used for accents
in all the languages.) The next screen asked for information about what
is bought, the price, the amount tendered, method of payment − still all
in German, as are the print commands, etc.
6.8
Having entered some information, I printed it out − a very basic till
receipt with my list of items and prices, which are totalled with the
amount of change calculated and my message from the first screen (thank
you). It was much the same as all those till receipts I have brought
back from various trips in recent years, except that at least the pupil
has to enter the items bought and understand approximately twenty items
of vocabulary which appear on the screen for each “customer”.
6.8
I began to plan tasks for my beginners using the printouts. One pupil
could produce one and the others ask questions to find out the prices.
We could produce them for different shops and compare prices.... but
then, I have a cupboard full of attractive, colourful supermarket
advertising leaflets from Germany, in class sets, surely it would be
better to use them!
6.8
I decided to go back to the program to see what else it has to offer
which the GCSE sets could use. I pressed “clear” to see what happened −
the chance to enter another list of items and prices! Since the computer
totals the prices and works out the change, you can’t even work on
complaints at the checkout, except by pretending that the wrong price
was entered.
6.8
There is no list of acceptable items, so there is no spellcheck, it
accepts complete nonsense − an item is any set of letters, a price any
number. There are no graphics, no variations on the basic layout and
content, nothing which will interest the pupils or give them a sense of
achievement, unless the teacher spends hours devising speaking tasks
based on till receipts.
6.8
Is it worth taking the whole class to the computer room to produce a
till receipt, or devising a flexible learning lesson with a variety of
small group activities so that this program can be used on the one
computer in my room? No, definitely not. I have some much better ideas
based on my wide selection of realia!
6.8
The program’s handbook suggests many sophisticated possibilities using
product files and bar codes which may be valid for teachers of
Technology, etc to develop. However, as a language teacher, I have not
yet found the time to work on these areas which seem to have little to
do with communication, verbal or written. In case I was too hasty in my
rejection of this program, I decided to use it with a class anyway. The
first pupil (below average year 9) got to the printing stage and said
“Now what?” − “That’s all it does”, I replied. He returned to his group
to do the structured oral tasks I had set up, showed them his till
receipt and the rest asked, “Wouldn’t it be quicker if we just wrote one
out miss?”  A
6.8
Hints and Tips
6.8
•    A5000 serial port problems − Several of you will have experienced
problems with dead or dying serial ports on their A5000s. Acorn are
aware of problems on early A5000 p.c.b.s and have a fix for them.
However, there are still some problems with static. Until Acorn find a
solution, the best we can do is take suitable precautions. (1) Use
braid-screened cable and connect the braid to the connector shells at
both ends. Also, never connect a cable when the machine at either end is
switched on. Always switch on both devices after the cable has been
connected.
6.8
•    Filer_CloseDir − On my 4Mb A3000, I like to open ‘related’ windows to
various programs and then close them again when finished, by using
Filer_OpenDir <pathname> <position> and Filer_ CloseDir <pathname>
respectively. I put these commands in the !Run file of !BJ10e Turbo
Drivers, one before the command to run !Runimage and the other after.
Readers should be warned that the command Filer_CloseDir clashes with
CC’s Turbo Driver (v2.09b) when using their Compression filing system
(CFS) (v1.17). If the Turbo driver is loaded first and then CFS, you
will find that, when trying to quit the Turbo Driver, it goes into an
error loop with SWI &43303 not known and only a <ctrl-break> or <reset>
will release the computer so you lose all unsaved files, etc. No problem
occurs when they are loaded in reverse order or when Filer_ CloseDir is
deleted from the !Run file of !BJ10e.
6.8
Does anyone know whether any other programs clash with Filer_CloseDir?
Brian Cocksedge, Midhurst.
6.8
•    Filofax printing − We had a letter printed on a piece of Filofax paper
which proved that you can do so with a Canon BJ200 with CC’s Turbo
Driver software using a custom page size of 95 × 165 mm. If you use
Impression (II or Junior), it shows the print area so it is easy then to
create a suitable frame and choose a small typeface (say, 8 pt) and view
the page at 200%.  Brian Stewart, Nairn.
6.8
•    Inkjet printing − When inkjet/bubblejet printouts need a relatively
large amount of ink, the paper can become a little wrinkled. To remove
the wrinkles, pass the paper through a laser printer by printing a blank
page. This calendars the paper to its original smooth finish.  George
Foot, Oxted.
6.8
•    Laser toner cartridges − The first sign you get that a laser toner
cartridge is about to run out is a line down the paper which is
obviously very low on toner. If you “rock the cartridge gently”, you may
be able to spread the toner out sufficiently to get a few more sheets
out of it. Try rocking first but then eventually, try shaking. Beware,
if you shake too soon, you can end up with toner all over the place!
Colin Singleton, Sheffield.
6.8
•    ShowPage confusion − It has been said within the pages of Archive that
“ShowPage does not work with RISC OS 3”. This is not entirely correct.
It will not work with RISC OS 3 fonts, i.e. those which have their own
kerning tables, and it will not work with RISC OS 3’s ROM fonts.
However, if you want to use ShowPage, all you need to do is to replace
the in-built fonts (Corpus, Homerton and Trinity) with ones within your
own font folder and included a FontRemove command to remove the ROM
fonts.  A
6.8

Hints and Tips
6.9
•    Anti-aliased fonts − I find that anti-aliasing on a standard
resolution monitor gives a far too blurred image. My hint is to use the
outline version of the System font at an unexpanded point size. A bit of
experimentation is needed to get it right if you habitually use the
screen at some scale other than one to one but when you get it right,
the result is a nice fast, clean and clear font on any DTP or
wordprocessing package while you work on the text. Remember to change
the font to your preferred version for the final print. I am afraid I
can’t remember where I obtained the Outline System font.
6.9
On my Pentoms game, I changed the transfer coefficients to sharpen up
the edges of my fancy fonts but that does not work well/at all on
RISC OS 3. Is there a way of getting a program to automatically check
which OS it is running under so that I can disable this facility where
it won’t work? Or is there another way of sharpening RISC OS 3 fonts?
Yes I know about turning off anti-aliasing but that is not quite the
same thing at all − bits of the letters are missed if you do that.
Simon Anthony, Nottingham.
6.9
• Backing up in one pass − The “Next” slot in the Task Manager can be
set from a !Boot file in RISC OS 3 to make backing up and copying of
full 800Kb discs occur in a single pass. Use the command
6.9
WimpSlot -next 864K
6.9
inside your !Boot file.  Barry Thompson,
6.9
Alternatively (for RISC OS 2 owners) you can allocate more memory to the
Next Task slot by issuing the Wimp_SlotSize SWI call from Basic. All
that you need is a Basic program containing the line:
6.9
SYS “Wimp_SlotSize”,-1,<slotsize>
6.9
where <slotsize> is in bytes (e.g. use 65536 for 64Kb.)
6.9
Simon Moy, NCS.
6.9
•    !Boot files are not strictly necessary within applications. They use
up disc space and slow down the opening of windows (especially over a
network). If an application’s !Boot file only contains the IconSprites
command, then this can be moved to the !Run file, and the !Boot file can
be deleted. The only effect is that the application will not show the
correct icon in the filer window until the application is actually run.
Keith Harja, Derby.
6.9
•    Gods passwords − On this month’s program disc is all you need for
getting the passwords for Gods. Andrew Ferguson, London N15.
6.9
•    Mono TIFF files − In reply to help for Ray Dawson in Archive 6.5 p18
for producing mono TIFF files from B&W scanned sprites. I have used
!MakeTiff version 2.00 by John Kortink to do exactly that. I examined
the resulting Tiff file using Graphics Workshop (under the emulator) and
Translator, and they both showed the converted sprite as a B&W image.
Chris Bass, Grimsby.
6.9
•    Symbols for Impression − When using Impression, or even more so
Junior, it can sometimes be tedious to find a desired symbol from a font
such as Dingbats. Therefore, I have created a number of text files which
can be dragged into any Impression document. The format for such files
is “® ”, without the quotes. This returns the type to its default size,
selects the desired font, inserts the character with ASCII code 174 and
resets the font to its default. N.B. The space at the end is essential.
This is so that one can do a <ctrl-right> to end of text. Without this
space, the font is not reset. I have given a number of pre-prepared ones
to Paul for inclusion on this month’s program disc.  Neil Walker,
Norwich.
6.9
You need to be very careful with this one. I tried it on Impression 2.19
and it gave an internal error and Impression bombed out! This was
because I didn’t have the symbol font in my system. I tried it on
version 2.17 and it complained politely that it didn’t have Symbol font
and let me carry on. I presume that the difference is because
Impression, from 2.18 onwards, has no limit to the number of fonts it
can handle so that part of the code must have been changed. Don’t say we
didn’t warn you! Ed.
6.9
•    Testing printers’ on-line status − Here is a way of testing whether
the printer is on- or off-line:
6.9
 10 *FX 3,10
6.9
 20 PRINT CHR$(0);
6.9
 30 a = INKEY(20)
6.9
 40 a = ADVAL(-4)
6.9
 50 *FX 3,0
6.9
 60 *FX 21,3
6.9
 70 IF a <> 1023 THEN
6.9
 80    PRINT “Printer not on line”
6.9
 90 ELSE
6.9
100   PRINT “Printer on line”
6.9
110 ENDIF
6.9
120 END
6.9
Paul Ribworth, Gloucester
6.9
•    Using high density discs − Until the advent of the A5000, all
Archimedes computers had an 800Kb 3½“ disc drive so the problems of
different disc densities didn’t really arise. With the introduction of
the new A3000 and A4000 series, 1.6Mb drives are becoming more common
and some users are trying to save a few pence by formatting normal
density discs to 1.6Mb.
6.9
Users of BBC and Master computers learned that there is no point in
paying extra for 80 track discs as the cheaper 40 track were actually
identical. The same is NOT true of 800Kb and 1.6Mb discs. The magnetic
coating on high density discs uses finer particles than the coating used
on standard (double density) types. This is designed to accept higher
signal levels without distortion. When the higher capacity format is
used, the drive automatically uses a stronger magnetic field to imprint
the data on the discs surface.
6.9
Sometimes, good quality normal density discs can accept this level but
it does represent a severe overload. Even if all appears well at the
time, the larger particles used in the normal density coating are
frequently unable to retain the high level of magnetism required so it
can leak away in time making part of the data unreadable.
6.9
Also, you should not use high density discs in a normal density drive if
they have previously been formatted to 1.44 or 1.6Mb. The lower flux
levels used for the normal format may not be able to completely erase
the earlier data. This can then ‘break through’ and corrupt data saved
later at the lower flux levels used with the smaller format. It may
sometimes take more than a month for this to happen. So don’t be fooled
into thinking that everything is OK just because all seems to be well
when you first format the disc. Many Archimedes users who have
‘borrowed’ preformatted 1.44Mb discs from work to use in their 800Kb
drives have experienced this problem.
6.9
There is no reason why you cannot use a NEW high density disc at 800Kb,
it will just be more expensive.
6.9
Because of this, it is obviously useless to do what used to be common
practice with the old 40/80 track discs, namely try to format them to 80
tracks and if it didn’t work use them for 40 track. Once you have
attempted to format an 800Kb disc to 1.6 or 1.44 Mb there is no going
back. I have had some success using a tape head demagnetiser to
completely wipe out data on 800Kb discs which had been formatted to
1.6Mb so that they could be reformatted to 800Kb, but not everyone has
one of these.
6.9
PC drives all have a sensor which detects the extra hole in a high
density disc and won’t allow you to specify the wrong format. For some
reason, Acorn didn’t implement this feature on the Archimedes although
the detector is present on all high density drives. It is therefore
possible, by accident, to format a normal density disc to 1.6Mb. On a
PC, to do this you need to drill a hole in the case of the disc so it
can’t happen unintentionally.
6.9
The most pernicious thing about this problem is that it can lead you
into a false sense of security. A normal density disc may format
perfectly to 1.6Mb and, for months, everything will appear to be well.
Then suddenly some of the data may become slightly corrupt, although the
disc may not actually appear to have any errors.
6.9
If your data is precious enough to be worth saving, it’s worth the extra
few pence that the proper discs cost. The only time that I have known
people use normal density discs successfully as high density was with
expensive branded discs costing about £1.30 each. As this is
considerably more than the price of unbranded high density discs, it
seems a foolish risk to take.  David Holden, Sydenham.  A
6.9
Hints and Tips
6.9
•    Anti-aliased fonts − I find that anti-aliasing on a standard resolution monitor gives a far too blurred image. My hint is to use the outline version of the System font at an unexpanded point size. A bit of experimentation is needed to get it right if you habitually use the screen at some scale other than one to one but when you get it right, the result is a nice fast, clean and clear font on any DTP or wordprocessing package while you work on the text. Remember to change the font to your preferred version for the final print. I am afraid I can’t remember where I obtained the Outline System font.
6.9
On my Pentoms game, I changed the transfer coefficients to sharpen up the edges of my fancy fonts but that does not work well/at all on RISC OS 3. Is there a way of getting a program to automatically check which OS it is running under so that I can disable this facility where it won’t work? Or is there another way of sharpening RISC OS 3 fonts? Yes I know about turning off anti-aliasing but that is not quite the same thing at all − bits of the letters are missed if you do that.  Simon Anthony, Nottingham.
6.9
• Backing up in one pass − The “Next” slot in the Task Manager can be set from a !Boot file in RISC OS 3 to make backing up and copying of full 800Kb discs occur in a single pass. Use the command
6.9
WimpSlot -next 864K
6.9
inside your !Boot file.  Barry Thompson,
6.9
Alternatively (for RISC OS 2 owners) you can allocate more memory to the Next Task slot by issuing the Wimp_SlotSize SWI call from Basic. All that you need is a Basic program containing the line:
6.9
SYS “Wimp_SlotSize”,-1,<slotsize>
6.9
where <slotsize> is in bytes (e.g. use 65536 for 64Kb.)
6.9
Simon Moy, NCS.
6.9
•    !Boot files are not strictly necessary within applications. They use up disc space and slow down the opening of windows (especially over a network). If an application’s !Boot file only contains the IconSprites command, then this can be moved to the !Run file, and the !Boot file can be deleted. The only effect is that the application will not show the correct icon in the filer window until the application is actually run.  Keith Harja, Derby.
6.9
•    Gods passwords − On this month’s program disc is all you need for getting the passwords for Gods. Andrew Ferguson, London N15.
6.9
•    Mono TIFF files − In reply to help for Ray Dawson in Archive 6.5 p18 for producing mono TIFF files from B&W scanned sprites. I have used !MakeTiff version 2.00 by John Kortink to do exactly that. I examined the resulting Tiff file using Graphics Workshop (under the emulator) and Translator, and they both showed the converted sprite as a B&W image.  Chris Bass, Grimsby.
6.9
•    Symbols for Impression − When using Impression, or even more so Junior, it can sometimes be tedious to find a desired symbol from a font such as Dingbats. Therefore, I have created a number of text files which can be dragged into any Impression document. The format for such files is “® ”, without the quotes. This returns the type to its default size, selects the desired font, inserts the character with ASCII code 174 and resets the font to its default. N.B. The space at the end is essential. This is so that one can do a <ctrl-right> to end of text. Without this space, the font is not reset. I have given a number of pre-prepared ones to Paul for inclusion on this month’s program disc.  Neil Walker, Norwich.
6.9
You need to be very careful with this one. I tried it on Impression 2.19 and it gave an internal error and Impression bombed out! This was because I didn’t have the symbol font in my system. I tried it on version 2.17 and it complained politely that it didn’t have Symbol font and let me carry on. I presume that the difference is because Impression, from 2.18 onwards, has no limit to the number of fonts it can handle so that part of the code must have been changed. Don’t say we didn’t warn you! Ed.
6.9
•    Testing printers’ on-line status − Here is a way of testing whether the printer is on- or off-line:
6.9
 10 *FX 3,10
6.9
 20 PRINT CHR$(0);
6.9
 30 a = INKEY(20)
6.9
 40 a = ADVAL(-4)
6.9
 50 *FX 3,0
6.9
 60 *FX 21,3
6.9
 70 IF a <> 1023 THEN
6.9
 80    PRINT “Printer not on line”
6.9
 90 ELSE
6.9
100   PRINT “Printer on line”
6.9
110 ENDIF
6.9
120 END
6.9
Paul Ribworth, Gloucester
6.9
•    Using high density discs − Until the advent of the A5000, all Archimedes computers had an 800Kb 3½“ disc drive so the problems of different disc densities didn’t really arise. With the introduction of the new A3000 and A4000 series, 1.6Mb drives are becoming more common and some users are trying to save a few pence by formatting normal density discs to 1.6Mb.
6.9
Users of BBC and Master computers learned that there is no point in paying extra for 80 track discs as the cheaper 40 track were actually identical. The same is NOT true of 800Kb and 1.6Mb discs. The magnetic coating on high density discs uses finer particles than the coating used on standard (double density) types. This is designed to accept higher signal levels without distortion. When the higher capacity format is used, the drive automatically uses a stronger magnetic field to imprint the data on the discs surface.
6.9
Sometimes, good quality normal density discs can accept this level but it does represent a severe overload. Even if all appears well at the time, the larger particles used in the normal density coating are frequently unable to retain the high level of magnetism required so it can leak away in time making part of the data unreadable.
6.9
Also, you should not use high density discs in a normal density drive if they have previously been formatted to 1.44 or 1.6Mb. The lower flux levels used for the normal format may not be able to completely erase the earlier data. This can then ‘break through’ and corrupt data saved later at the lower flux levels used with the smaller format. It may sometimes take more than a month for this to happen. So don’t be fooled into thinking that everything is OK just because all seems to be well when you first format the disc. Many Archimedes users who have ‘borrowed’ preformatted 1.44Mb discs from work to use in their 800Kb drives have experienced this problem.
6.9
There is no reason why you cannot use a NEW high density disc at 800Kb, it will just be more expensive.
6.9
Because of this, it is obviously useless to do what used to be common practice with the old 40/80 track discs, namely try to format them to 80 tracks and if it didn’t work use them for 40 track. Once you have attempted to format an 800Kb disc to 1.6 or 1.44 Mb there is no going back. I have had some success using a tape head demagnetiser to completely wipe out data on 800Kb discs which had been formatted to 1.6Mb so that they could be reformatted to 800Kb, but not everyone has one of these.
6.9
PC drives all have a sensor which detects the extra hole in a high density disc and won’t allow you to specify the wrong format. For some reason, Acorn didn’t implement this feature on the Archimedes although the detector is present on all high density drives. It is therefore possible, by accident, to format a normal density disc to 1.6Mb. On a PC, to do this you need to drill a hole in the case of the disc so it can’t happen unintentionally.
6.9
The most pernicious thing about this problem is that it can lead you into a false sense of security. A normal density disc may format perfectly to 1.6Mb and, for months, everything will appear to be well. Then suddenly some of the data may become slightly corrupt, although the disc may not actually appear to have any errors.
6.9
If your data is precious enough to be worth saving, it’s worth the extra few pence that the proper discs cost. The only time that I have known people use normal density discs successfully as high density was with expensive branded discs costing about £1.30 each. As this is considerably more than the price of unbranded high density discs, it seems a foolish risk to take.  David Holden, Sydenham.  A
6.9
Using RISC OS 3.10
6.9
Hugh Eagle
6.9
More encouragement for the faint-hearted!
6.9
Stuart Bell says: “Having read the RISC OS Column for the past few months, I approached the upgrade with caution, afraid of all sorts of installation and compatibility problems. I guess that by the nature of the things, it’s those people who encounter problems who write in, thereby giving a false impression of the true situation. So, if you’re still with RISC OS 2 and thinking about upgrading, be encouraged. Installing RISC OS 3 on my A310, including removing the main board and installing the IFEL ROM carrier board took about an hour from start to finish, and I experienced no unforeseen problems.
6.9
“Having been warned of problems with Compression 1.1, I de-compressed all files first. I await the free upgrade from CC. There is the reported problem with early versions of Font-FX and that’s in hand.
6.9
“A couple of warnings, though: As well as powering up with <Delete> depressed, also make sure that you’ve re-connected the keyboard first. (Don’t laugh – it could have been you! That cost me several minutes of panic.) Secondly, alter the Font Size figure before loading your first Impression document, which otherwise can take several minutes to load with a 32Kb buffer.”
6.9
• Avie Electronics 4Mb A310 upgrade − The following information has come from Avie: LK12 should be altered to 1-3 and 2-4. Issue 1 PCBs do not have LK12 and require modification as follows:
6.9
Option 1: pins 2 and 24 on all of the ROMs are to be swapped. Remove pins 2 and 24 from each ROM socket by bending them up. Join pin 24 on all ROMs to GND. Join pin 2 on all ROMs to LA18 on IC28, pin 18.
6.9
Option 2: remove pin 2 from each stamped socket on PCB and strap a wire between the pins and LA18. Likewise, pin 24 can be strapped to GND after removing the pins in the stamped sockets.
6.9
Hardware problems
6.9
Watford Electronics ARM3 − When Colin Wood fitted a Watford Electronics ARM3 upgrade to his A410 it didn’t work. “The screen flashed red/blue, but no beep. Just dead.” As he had recently fitted RISC OS 3, he decided to replace the RISC OS 2 chips and everything seemed OK, but when he put the RISC OS 3 chips back, again it didn’t work. Watford’s help line couldn’t help, so he returned the ARM3 and since they were (again) out of stock he cancelled his order and got most (but not all, apparently) of his money back. Another ARM3, from CJE Micro’s, worked first time.
6.9
Watford A310 RAM upgrades − Watford Electronics say that they are aware that their early A305/310 RAM upgrades do not work with RISC OS 3 but they do not have any solution. They simply recommend that you buy one of their newer RAM upgrades instead because they do work with RISC OS 3. Fortunately, IFEL have stepped into the breach. They can offer new PAL chips to plug into the Watford boards that will solve the problem. These PAL chips are £3 +VAT each from IFEL and you need to state whether you have the 2Mb or the 4Mb version. (This refers to the upgrades that consisted of two PCBs with a ribbon cable along the RHS of the machine.)
6.9
Program problems
6.9
QuicKey clash − J. Nichols reports that, if QuicKey is loaded from within a !Boot file, it can remove the *Opt 4,2 status from the booted disc resulting in a disc which no longer runs a !Boot file.
6.9
!Edit – Following last month’s reports of a “bug” in Edit we now have the following from Tom Rank: “There seemed to be a rare problem with Edit the other day; my text (exported from Ovation) was displayed in a most corrupted way, with a large amount of blank screen to start with and then disintegrating words. I thought it was a problem with Ovation. However, when I tried to show it to John Wallace of Risc Developments at BETT, the file displayed without problems, as it did back on my own machine the next day. Any idea why?”
6.9
Fun School 4 (for the under 5s) – Tom Rank found that this ran under RISC OS 3.00 but not 3.10, but Europress Software supplied an updated disc very promptly which now works.
6.9
Freddy Teddy – Topologika supplied a new version to Tom Rank on the spot at BETT!
6.9
Ovation – Tom Rank says you are urged to obtain 1.37S (dated 17 January 1993). This overcomes the saving with <F3> problem – which really was a problem, so do upgrade. They have also restored the ‘select all’ shortcut − but only if you use the right-hand <Ctrl> key along with A (to avoid the problem of catching <Ctrl-A> instead of <Ctrl-S> or just <A>).
6.9
This version also contains three new printing options, not documented: Fit, Centre and Registration marks. These should be self explanatory, but Tom is not quite sure about them!
6.9
Break147 and SuperPool – D. S. Allen got these to work using the information provided by David Holden in the April column. He writes: “In Break147 there is one Basic program called !Snook which contains the line
6.9
 90SYS 26,64,64.
6.9
A REM statement in front of the SYS will cure the problem and the program will then run under RISC OS 3.1.
6.9
“Similarly, in SuperPool there is a Basic program !Pool in the Data directory in which line 60 has the same SYS 26,64,64 statement.
6.9
“Note that to make these changes, people will require the read/write version of Mark Smith’s excellent ArcFS program.” (Or, presumably, an equivalent utility such as SparkFS.)
6.9
Superior Golf / Construction Set – We reported in Archive 6.7 p43 that Dave Wilcox had found that these did not work in RISC OS 3.10. Rob Brown, however, says he has had no problems with them.
6.9
Corruption – Rob Brown says that this can be made to work by modifying its !Run file to kill the RISC OS 3 Shared C Library module and load the very old version of the module supplied with the game. To do this, include at the beginning of the !Run file the line:
6.9
RMKill SharedCLibrary
6.9
and then add at the end of the file the following two lines:
6.9
RMKill SharedCLibrary
6.9
RMLoad SharedCLibrary
6.9
to remove the old version and reload the RISC OS 3 version of the module. He points out that it may well be advisable to Quit any other applications that may be using the RISC OS 3 Shared C Library module before running the game.
6.9
Freddy’s Folly – Rob Brown advises that this game contains a !Run file (in Basic) which includes an automatic ‘auto-configuration’ of the computer’s memory as well as setting up a number of initial program parameters. The various auto-configuration elements should be deleted from this file. A further complication is that the disc is copy protected, but it should be possible to rename the original !Run file (to say !RunOLD) and re-save a new !Run file containing only the initial program parameters. He also advises that the game assumes that the default disc drive is 0. For users with a hard disc this can easily be set by typing:
6.9
    Drive 0
6.9
at the command line (accessed by pressing function key <F12>) before starting the game. Alternatively, the game’s !Run file can be appropriately amended by including the “<Obey$Dir>” command.
6.9
NOTE: the advice on Corruption and Freddy’s Folly is included in the program compatibility chart which Rob Brown has compiled and which has been included on the Archive monthly program disc. He will welcome any information which will help him to make the chart as comprehensive and up-to-date as possible. His address is “Valtanee”, Brighton Road, Lower Kingswood, Tadworth, Surrey, KT20 6UP.
6.9
Aldebaran – (see Archive 6.8 p23) Atle Mjelde Bårdholt has contacted the author and has learned that, on his A310 with IFEL’s 2Mb upgrade, he has to add 8Kb to the WimpSlot values in the !Aldebaran.!Run file (from 624Kb to 632Kb). Then everything works fine. He says this might be due to the RAM upgrade: Aldebaran seems to work fine on any other Archimedes with RISC OS 3.1. He adds a final bit of advice: “Get a copy!”
6.9
Printing
6.9
Printing to file
6.9
Tom Rank has found that the hint from Simon Moy in Archive 6.2.7 about editing the number of text lines seems redundant under RISC OS 3.10, and that printing to a file (at least with the Epson LQ-860 driver) resulted in a printout which aborted with an “address exception at &038B5BE0”, a frozen printer and an incomplete document (the last few millimetres were lost). This is particularly irritating as printing to a file is the only way, apart from buying a direct laser printer or a CC Turbo Driver, to avoid the long wait for print out. (How about using the PROdriver shell for use with Acorn drivers? See the review on page 43. Ed.)
6.9
Using fanfold paper
6.9
Tom Rank comments: “Feeding single sheets into a printer soon becomes a pain, but sheetfeeders come expensive, so A4 fanfold paper seemed a good idea (even though it’s about four or five times more expensive than better quality photocopy paper at a discount stationer!). However, with RISC OS 3.00, I could never get the settings right, so the print always crept up or down the page until it crossed the fold. With RISC OS 3.10, I’ve finally got it to work, whether by fluke I’m not sure. Using the LQ-860 printer driver with my Epson LQ-550 printer, the correct settings from the paper sizes menu are: top margin 8mm and bottom margin 15mm; save these with a name like ‘A4 Fanfold’ and the margins can be automatically set each time. These settings seem to work OK with single sheets too.”
6.9
Filing Systems
6.9
Beebug 5¼“ disc interface
6.9
Stuart Bell writes: “I got my Beebug 5¼“ disc interface (date 1989) to work by setting the step rate for an 80 track double-sided drive to 12ms (lower values may well work; I’m just cautious), and whereas RISC OS 2 worked with “Ready” (switch 6 on the interface) supplied by the on-board PAL, RISC OS 3 works more reliably with it passed through from the drive.”
6.9
Dismounting between filing systems
6.9
According to Mike Sawle, applications which frequently switch between filing systems seem to require the *DISMOUNT command before selecting an alternative filing system in order to ensure correct execution. Common error messages are Ambiguous Disc Name and Bad Error Block.
6.9
Miscellaneous Hints & Tips
6.9
•    Faster copying, etc. – Rob Davison recommends that if there is nothing else you want to do when copying files, then click <Menu> over the filer action window and choose ‘Faster’.
6.9
•    In the midnight hour – Atle Mjelde Bårdholt found that, whenever he switched his computer on between midnight and 1 a.m., a day was added to the internal clock. He then found that this seemed to be caused by the presence in his !Boot file of the command *Set Sys$Year 1993 (which he had included because he has some programs that set the year back to 1988). When this command is removed, the midnight hour problem goes away. Does anyone know the reason?
6.9
•    Fan filters and the A5000 – Tom Rank found that, although magazines mention the need to change the filter from time to time, there was no mention of how to do this and on fitting the RISC OS 3.10 upgrade he could see no filter anyway. Is there one, he asks? The Acorn representative at BETT told him that the design of the machine made it superfluous; he couldn’t quite follow why, but it was something to do with the direction of airflow.
6.9
•    Flippin’ objects in Draw – Having travelled all the way from Derbyshire to the Acorn User Show, Tom Rank “asked the man on one of the other magazines’ stall if there was an equivalent in Draw to the ‘flip’ option in Paint, as it’s sometimes useful to turn a picture round to face the other way. The expert, although helpful about other items, couldn’t tell me, but my son revealed the following tip, which he said he thought was obvious. Must talk to him more often....
6.9
“To flip an object in Draw, select it and, from the ‘transform’ menu (RISC OS 3), set x scale (or y scale or both) to −1 (minus 1). This also works in the RISC OS 2 version of Draw but there the ‘transform’ option is chosen from the Select menu.”
6.9
•    Lost lines in Edit − Eric Ayers has found that double-clicking on large text files to load them into Edit has caused several lines to be missing from the middle of the text. Apparently, this does not happen if he drags the files to the Edit icon on the icon bar. Any ideas?
6.9
•    More fonts in RISC-OS 3 − Gary Atkinson of KudlianSoft recommends the following additions to the !Fonts.!Boot and !Fonts.!Run files to enable one or more separate font directories to be added to and removed from the font path simply by double-clicking on the fonts application directory.
6.9
The !Fonts.!Boot file should be amended as follows:
6.9
| Boot file for !Fonts.
6.9
| Nothing happens if you put this in | a RISC OS 2 machine except that
6.9
| the !Fonts sprite is set up − this | means that, when a RISC OS2 !Fonts
6.9
| directory is seen the !Boot file
6.9
| will not be run; double clicking
6.9
| on the appropriate !Fonts will
6.9
| force installation.
6.9
|
6.9
| Take care when merging RISC OS 2
6.9
| fonts with these (RISC OS 3) fonts
6.9
| - many fonts are contained in the
6.9
| RISC OS 3 ROM and thus do not
6.9
| appear in !Fonts; normally you
6.9
| should not copy the corresponding
6.9
| RISC OS 2 fonts!
6.9
Iconsprites <Obey$Dir>.!Sprites
6.9
RMEnsure FontManager 2.89 Obey
6.9
| The next command boots up your
6.9
| normal font directory
6.9
FontInstall <Obey$Dir>.
6.9
SetEval MoreFonts 0
6.9
The !Fonts.!Run file should be amended as follows:
6.9
| !Run file for !Fonts,
6.9
| version 1.05 (20-Aug-91)
6.9
RMEnsure FontManager 2.89 Error These fonts are for use with
6.9
 RISC OS 2.12 or later
6.9
|Obey <Obey$Dir>.!Boot
6.9
| The extra *FontInstall means that,
6.9
| if you double-click the Fonts
6.9
| directory, all such directories
6.9
| will be re-searched for any new
6.9
| fonts.
6.9
SetEval MoreFonts MoreFonts EOR 1
6.9
IF MoreFonts=1 THEN FontInstall <Obey$Dir>. ELSE FontRemove <Obey$Dir>.
6.9
|FontInstall
6.9
The !Boot file is only run on the first time that the filer sees the !Fonts application, hence the initialisation of MoreFonts to zero. Every time the directory is double-clicked, the !Run file will toggle the value of this variable between 0 and 1. This is done in the line:
6.9
SetEval MoreFonts MoreFonts EOR 1
6.9
If the value is 1, the FontInstall command adds the new directory to the font path and initialises all of the fonts. If the value is 0, the FontRemove command removes the temporary font directory from the font path.
6.9
If you require more than one spare directory, you will need to use different variables, such as MoreFonts2 or FancyFonts.
6.9
In practice, it is better to have a directory called fonts inside your !Fonts application which contains your preferred normal fonts at switch-on, and a directory called spare (for example) which contains the set of fonts that you wish to switch in and out. Then in your !Boot file, the FontInstall command becomes
6.9
FontInstall <Obey$Dir>.fonts.
6.9
and the IF...THEN statement in the !Run file has <Obey$Dir>.spare. in place of the two occurrences of <Obey$Dir>. .
6.9
Miscellaneous Comments
6.9
FontEd
6.9
When Tom Rank spoke to an Acorn representative he was told that they were not contemplating the release of FontEd and that it is NOT public domain. An early version got there by mistake via SID and it really ought not, he said, to be distributed by PD libraries.
6.9
The Acorn representative acknowledged that Acorn were unlikely to do anything about this situation, as it was “only an early version and won’t work on RISC OS 3”. According to Tom this is partially true; a friend of his has had problems with version 0.28, but at least some of the time it runs on an A5000. However, he observes that it is a pity that Acorn actually seem to be discouraging anyone else from working on fonts. (There is a version of FontEd on Careware 7, released in May 1990, but it has been updated since then. Acorn haven’t objected, presumably because of the charitable aspect. Ed.)
6.9
Screen blanking during floppy disc access
6.9
Stuart Bell writes: “Many things are very much more elegant in RISC OS 3 but with one notable exception. I couldn’t believe my eyes when I saw what happens to the screen during floppy disc accessing. Whether my ARM3 cache is dis- or enabled, in mode 24 the screen either flashes or is blanked when loading, saving or formatting. If RISC OS 2 didn’t need to do this, why is it allowed now? (It did, in fact, happen in mode 21 under RISC OS 2 – HE.) Some re-coding is needed, even if the resultant code is a little slower. More significantly, who was the Acorn person in authority who thought that it is acceptable in 1993 to produce a computer which professes to be a ‘world-beater’ which, when using floppy discs, looks like the Sinclair Spectrum (or was it the ZX81?) reading from cassette tape circa 1983? This just looks so tacky. I know that serious use demands a hard disc, but first impressions count for a lot, and a lot of demonstrations will be on floppy-based A3010’s. This is not good enough, Acorn.”
6.9
Basic program compaction errors
6.9
R. W. Darlington writes: “I am writing on a matter of utmost urgency.
6.9
“There seems to be a growing urge amongst programmers to use a program compactor on Basic RunImage files. Unfortunately, I know of no Basic compactor program that does not exhibit a certain disastrous fault.
6.9
“If, for instance, a Basic program contains any floating point or pseudo variables followed by a Basic keyword or statement, then when it is compacted and the space between the variable and the keyword is removed, that line will generate an error.
6.9
“For instance, if the program contained the line
6.9
10 IF TRUE THEN VDU7
6.9
which when compacted becomes
6.9
10IFTRUETHENVDU7
6.9
the lack of space between TRUE and THEN will generate an error.
6.9
“The same is true for any line containing a floating point variable:
6.9
10 ha=TRUE
6.9
20 IF    ha    THEN    VDU7
6.9
will crash when compacted because of a lack of space between ha and THEN.
6.9
“I have come across numerous examples of RunImage files which have these errors in them. One of them is Acorn’s new RISC OS 3 !SciCalc. Others are new versions of Translator and Creator.
6.9
“Please can you draw Archive readers’ attention to these problems as soon as possible? Or, if any reader knows of a Basic compactor that does not produce these errors, please can they let us know where it may be obtained?”
6.9
Matters Arising
6.9
Setting Caps Lock (*FX 202)
6.9
Unfortunately, the advice I gave last month about using *FX 202 may have been a bit confusing because, after explaining that Peter Young wanted a command to turn Caps Lock on, I then described how to turn it off! The command to turn Caps Lock on, leaving the other settings unchanged, is *FX 202,0,239.
6.9
The shortened form *FX 202,32 should turn Caps Lock on, but may have side effects.
6.9
Note also that, to make sure that the LED indicators are correctly updated after a *FX 202 command, it is a good idea to follow it with the simple command *FX 118.
6.9
The bits of the keyboard status byte have the following meanings:
6.9
    bit    value    meaning
6.9
    0 (1)    0    Alt off
6.9
        1    Alt on
6.9
    1 (2)    0    Scroll Lock off
6.9
        1    Scroll Lock on
6.9
    2 (4)    0    Num Lock on
6.9
        1    Num Lock off
6.9
    3 (8)    0    Shift off
6.9
        1    Shift on
6.9
    4 (16)    0    Caps Lock on
6.9
        1    Caps Lock off
6.9
    5 (32)        (“Normally set”)
6.9
    6 (64)    0    Ctrl off
6.9
        1    Ctrl on
6.9
    7 (128)    0    Shift enable off
6.9
        1    Shift enable on
6.9
(This information is taken from the RISC­OS 2 Programmers’ Reference Manual.)
6.9
Boot-up options (Archive 6.8 p26)
6.9
Peter Young has suggested that it might be better to use INKEY$ with a delay of, say, 5 to 10 seconds instead of GET$ so that, if nobody presses a key, the computer will continue with the boot file after a short delay.
6.9
Unexpected messages from IDE filing system (Archive 6.8 p24)
6.9
Peter Young (who has a Beebug IDE interface) has confirmed that, even after creating space for a directory cache, he still gets unexpected messages unless he uses full pathnames. For instance, if he starts with the commands
6.9
*DIR IDEFS::IDEDisc4.$.BASIC.B
6.9
*CAT
6.9
he gets a catalogue of the subdirectory ‘B’ as expected. However, if he uses the abbreviated pathname instead:
6.9
*DIR $.BASIC.B
6.9
*CAT
6.9
he gets a catalogue of the directory ‘BASIC’ even though the heading of the directory listing correctly says that the CSD is IDEFS::IDEDisc4.$. BASIC.B. The computer’s confusion is further demonstrated by the fact that if he then enters:
6.9
*DIR ^
6.9
*CAT
6.9
it replies Directory ‘@’ not found.
6.9
‘Interrupt’ module (Archive 6.6 p55)
6.9
In March, we reported M. P. Sawle’s request for help in getting the Interrupt module (Archive Shareware disc 6) working under RISC OS 3, but since then neither he nor we have heard anything. Does no-one else use this module? Can its author help?
6.9
Using RISC OS 3.10
6.9
Hugh Eagle
6.9
More encouragement for the faint-hearted!
6.9
Stuart Bell says: “Having read the RISC OS Column for the past few
months, I approached the upgrade with caution, afraid of all sorts of
installation and compatibility problems. I guess that by the nature of
the things, it’s those people who encounter problems who write in,
thereby giving a false impression of the true situation. So, if you’re
still with RISC OS 2 and thinking about upgrading, be encouraged.
Installing RISC OS 3 on my A310, including removing the main board and
installing the IFEL ROM carrier board took about an hour from start to
finish, and I experienced no unforeseen problems.
6.9
“Having been warned of problems with Compression 1.1, I de-compressed
all files first. I await the free upgrade from CC. There is the reported
problem with early versions of Font-FX and that’s in hand.
6.9
“A couple of warnings, though: As well as powering up with <Delete>
depressed, also make sure that you’ve re-connected the keyboard first.
(Don’t laugh – it could have been you! That cost me several minutes of
panic.) Secondly, alter the Font Size figure before loading your first
Impression document, which otherwise can take several minutes to load
with a 32Kb buffer.”
6.9
• Avie Electronics 4Mb A310 upgrade − The following information has come
from Avie: LK12 should be altered to 1-3 and 2-4. Issue 1 PCBs do not
have LK12 and require modification as follows:
6.9
Option 1: pins 2 and 24 on all of the ROMs are to be swapped. Remove
pins 2 and 24 from each ROM socket by bending them up. Join pin 24 on
all ROMs to GND. Join pin 2 on all ROMs to LA18 on IC28, pin 18.
6.9
Option 2: remove pin 2 from each stamped socket on PCB and strap a wire
between the pins and LA18. Likewise, pin 24 can be strapped to GND after
removing the pins in the stamped sockets.
6.9
Hardware problems
6.9
Watford Electronics ARM3 − When Colin Wood fitted a Watford Electronics
ARM3 upgrade to his A410 it didn’t work. “The screen flashed red/blue,
but no beep. Just dead.” As he had recently fitted RISC OS 3, he decided
to replace the RISC OS 2 chips and everything seemed OK, but when he put
the RISC OS 3 chips back, again it didn’t work. Watford’s help line
couldn’t help, so he returned the ARM3 and since they were (again) out
of stock he cancelled his order and got most (but not all, apparently)
of his money back. Another ARM3, from CJE Micro’s, worked first time.
6.9
Watford A310 RAM upgrades − Watford Electronics say that they are aware
that their early A305/310 RAM upgrades do not work with RISC OS 3 but
they do not have any solution. They simply recommend that you buy one of
their newer RAM upgrades instead because they do work with RISC OS 3.
Fortunately, IFEL have stepped into the breach. They can offer new PAL
chips to plug into the Watford boards that will solve the problem. These
PAL chips are £3 +VAT each from IFEL and you need to state whether you
have the 2Mb or the 4Mb version. (This refers to the upgrades that
consisted of two PCBs with a ribbon cable along the RHS of the machine.)
6.9
Program problems
6.9
QuicKey clash − J. Nichols reports that, if QuicKey is loaded from
within a !Boot file, it can remove the *Opt 4,2 status from the booted
disc resulting in a disc which no longer runs a !Boot file.
6.9
!Edit – Following last month’s reports of a “bug” in Edit we now have
the following from Tom Rank: “There seemed to be a rare problem with
Edit the other day; my text (exported from Ovation) was displayed in a
most corrupted way, with a large amount of blank screen to start with
and then disintegrating words. I thought it was a problem with Ovation.
However, when I tried to show it to John Wallace of Risc Developments at
BETT, the file displayed without problems, as it did back on my own
machine the next day. Any idea why?”
6.9
Fun School 4 (for the under 5s) – Tom Rank found that this ran under
RISC OS 3.00 but not 3.10, but Europress Software supplied an updated
disc very promptly which now works.
6.9
Freddy Teddy – Topologika supplied a new version to Tom Rank on the spot
at BETT!
6.9
Ovation – Tom Rank says you are urged to obtain 1.37S (dated 17 January
1993). This overcomes the saving with <F3> problem – which really was a
problem, so do upgrade. They have also restored the ‘select all’
shortcut − but only if you use the right-hand <Ctrl> key along with A
(to avoid the problem of catching <Ctrl-A> instead of <Ctrl-S> or just
<A>).
6.9
This version also contains three new printing options, not documented:
Fit, Centre and Registration marks. These should be self explanatory,
but Tom is not quite sure about them!
6.9
Break147 and SuperPool – D. S. Allen got these to work using the
information provided by David Holden in the April column. He writes: “In
Break147 there is one Basic program called !Snook which contains the
line
6.9
 90SYS 26,64,64.
6.9
A REM statement in front of the SYS will cure the problem and the
program will then run under RISC OS 3.1.
6.9
“Similarly, in SuperPool there is a Basic program !Pool in the Data
directory in which line 60 has the same SYS 26,64,64 statement.
6.9
“Note that to make these changes, people will require the read/write
version of Mark Smith’s excellent ArcFS program.” (Or, presumably, an
equivalent utility such as SparkFS.)
6.9
Superior Golf / Construction Set – We reported in Archive 6.7 p43 that
Dave Wilcox had found that these did not work in RISC OS 3.10. Rob
Brown, however, says he has had no problems with them.
6.9
Corruption – Rob Brown says that this can be made to work by modifying
its !Run file to kill the RISC OS 3 Shared C Library module and load the
very old version of the module supplied with the game. To do this,
include at the beginning of the !Run file the line:
6.9
RMKill SharedCLibrary
6.9
and then add at the end of the file the following two lines:
6.9
RMKill SharedCLibrary
6.9
RMLoad SharedCLibrary
6.9
to remove the old version and reload the RISC OS 3 version of the
module. He points out that it may well be advisable to Quit any other
applications that may be using the RISC OS 3 Shared C Library module
before running the game.
6.9
Freddy’s Folly – Rob Brown advises that this game contains a !Run file
(in Basic) which includes an automatic ‘auto-configuration’ of the
computer’s memory as well as setting up a number of initial program
parameters. The various auto-configuration elements should be deleted
from this file. A further complication is that the disc is copy
protected, but it should be possible to rename the original !Run file
(to say !RunOLD) and re-save a new !Run file containing only the initial
program parameters. He also advises that the game assumes that the
default disc drive is 0. For users with a hard disc this can easily be
set by typing:
6.9
    Drive 0
6.9
at the command line (accessed by pressing function key <F12>) before
starting the game. Alternatively, the game’s !Run file can be
appropriately amended by including the “<Obey$Dir>” command.
6.9
NOTE: the advice on Corruption and Freddy’s Folly is included in the
program compatibility chart which Rob Brown has compiled and which has
been included on the Archive monthly program disc. He will welcome any
information which will help him to make the chart as comprehensive and
up-to-date as possible. His address is “Valtanee”, Brighton Road, Lower
Kingswood, Tadworth, Surrey, KT20 6UP.
6.9
Aldebaran – (see Archive 6.8 p23) Atle Mjelde Bårdholt has contacted the
author and has learned that, on his A310 with IFEL’s 2Mb upgrade, he has
to add 8Kb to the WimpSlot values in the !Aldebaran.!Run file (from
624Kb to 632Kb). Then everything works fine. He says this might be due
to the RAM upgrade: Aldebaran seems to work fine on any other Archimedes
with RISC OS 3.1. He adds a final bit of advice: “Get a copy!”
6.9
Printing
6.9
Printing to file
6.9
Tom Rank has found that the hint from Simon Moy in Archive 6.2.7 about
editing the number of text lines seems redundant under RISC OS 3.10, and
that printing to a file (at least with the Epson LQ-860 driver) resulted
in a printout which aborted with an “address exception at &038B5BE0”, a
frozen printer and an incomplete document (the last few millimetres were
lost). This is particularly irritating as printing to a file is the only
way, apart from buying a direct laser printer or a CC Turbo Driver, to
avoid the long wait for print out. (How about using the PROdriver shell
for use with Acorn drivers? See the review on page 43. Ed.)
6.9
Using fanfold paper
6.9
Tom Rank comments: “Feeding single sheets into a printer soon becomes a
pain, but sheetfeeders come expensive, so A4 fanfold paper seemed a good
idea (even though it’s about four or five times more expensive than
better quality photocopy paper at a discount stationer!). However, with
RISC OS 3.00, I could never get the settings right, so the print always
crept up or down the page until it crossed the fold. With RISC OS 3.10,
I’ve finally got it to work, whether by fluke I’m not sure. Using the
LQ-860 printer driver with my Epson LQ-550 printer, the correct settings
from the paper sizes menu are: top margin 8mm and bottom margin 15mm;
save these with a name like ‘A4 Fanfold’ and the margins can be
automatically set each time. These settings seem to work OK with single
sheets too.”
6.9
Filing Systems
6.9
Beebug 5¼“ disc interface
6.9
Stuart Bell writes: “I got my Beebug 5¼“ disc interface (date 1989) to
work by setting the step rate for an 80 track double-sided drive to 12ms
(lower values may well work; I’m just cautious), and whereas RISC OS 2
worked with “Ready” (switch 6 on the interface) supplied by the on-board
PAL, RISC OS 3 works more reliably with it passed through from the
drive.”
6.9
Dismounting between filing systems
6.9
According to Mike Sawle, applications which frequently switch between
filing systems seem to require the *DISMOUNT command before selecting an
alternative filing system in order to ensure correct execution. Common
error messages are Ambiguous Disc Name and Bad Error Block.
6.9
Miscellaneous Hints & Tips
6.9
•    Faster copying, etc. – Rob Davison recommends that if there is nothing
else you want to do when copying files, then click <Menu> over the filer
action window and choose ‘Faster’.
6.9
•    In the midnight hour – Atle Mjelde Bårdholt found that, whenever he
switched his computer on between midnight and 1 a.m., a day was added to
the internal clock. He then found that this seemed to be caused by the
presence in his !Boot file of the command *Set Sys$Year 1993 (which he
had included because he has some programs that set the year back to
1988). When this command is removed, the midnight hour problem goes
away. Does anyone know the reason?
6.9
•    Fan filters and the A5000 – Tom Rank found that, although magazines
mention the need to change the filter from time to time, there was no
mention of how to do this and on fitting the RISC OS 3.10 upgrade he
could see no filter anyway. Is there one, he asks? The Acorn
representative at BETT told him that the design of the machine made it
superfluous; he couldn’t quite follow why, but it was something to do
with the direction of airflow.
6.9
•    Flippin’ objects in Draw – Having travelled all the way from
Derbyshire to the Acorn User Show, Tom Rank “asked the man on one of the
other magazines’ stall if there was an equivalent in Draw to the ‘flip’
option in Paint, as it’s sometimes useful to turn a picture round to
face the other way. The expert, although helpful about other items,
couldn’t tell me, but my son revealed the following tip, which he said
he thought was obvious. Must talk to him more often....
6.9
“To flip an object in Draw, select it and, from the ‘transform’ menu
(RISC OS 3), set x scale (or y scale or both) to −1 (minus 1). This also
works in the RISC OS 2 version of Draw but there the ‘transform’ option
is chosen from the Select menu.”
6.9
•    Lost lines in Edit − Eric Ayers has found that double-clicking on
large text files to load them into Edit has caused several lines to be
missing from the middle of the text. Apparently, this does not happen if
he drags the files to the Edit icon on the icon bar. Any ideas?
6.9
•    More fonts in RISC-OS 3 − Gary Atkinson of KudlianSoft recommends the
following additions to the !Fonts.!Boot and !Fonts.!Run files to enable
one or more separate font directories to be added to and removed from
the font path simply by double-clicking on the fonts application
directory.
6.9
The !Fonts.!Boot file should be amended as follows:
6.9
| Boot file for !Fonts.
6.9
| Nothing happens if you put this in | a RISC OS 2 machine except that
6.9
| the !Fonts sprite is set up − this | means that, when a RISC OS2
!Fonts
6.9
| directory is seen the !Boot file
6.9
| will not be run; double clicking
6.9
| on the appropriate !Fonts will
6.9
| force installation.
6.9
|
6.9
| Take care when merging RISC OS 2
6.9
| fonts with these (RISC OS 3) fonts
6.9
| - many fonts are contained in the
6.9
| RISC OS 3 ROM and thus do not
6.9
| appear in !Fonts; normally you
6.9
| should not copy the corresponding
6.9
| RISC OS 2 fonts!
6.9
Iconsprites <Obey$Dir>.!Sprites
6.9
RMEnsure FontManager 2.89 Obey
6.9
| The next command boots up your
6.9
| normal font directory
6.9
FontInstall <Obey$Dir>.
6.9
SetEval MoreFonts 0
6.9
The !Fonts.!Run file should be amended as follows:
6.9
| !Run file for !Fonts,
6.9
| version 1.05 (20-Aug-91)
6.9
RMEnsure FontManager 2.89 Error These fonts are for use with
6.9
 RISC OS 2.12 or later
6.9
|Obey <Obey$Dir>.!Boot
6.9
| The extra *FontInstall means that,
6.9
| if you double-click the Fonts
6.9
| directory, all such directories
6.9
| will be re-searched for any new
6.9
| fonts.
6.9
SetEval MoreFonts MoreFonts EOR 1
6.9
IF MoreFonts=1 THEN FontInstall <Obey$Dir>. ELSE FontRemove <Obey$Dir>.
6.9
|FontInstall
6.9
The !Boot file is only run on the first time that the filer sees the
!Fonts application, hence the initialisation of MoreFonts to zero. Every
time the directory is double-clicked, the !Run file will toggle the
value of this variable between 0 and 1. This is done in the line:
6.9
SetEval MoreFonts MoreFonts EOR 1
6.9
If the value is 1, the FontInstall command adds the new directory to the
font path and initialises all of the fonts. If the value is 0, the
FontRemove command removes the temporary font directory from the font
path.
6.9
If you require more than one spare directory, you will need to use
different variables, such as MoreFonts2 or FancyFonts.
6.9
In practice, it is better to have a directory called fonts inside your
!Fonts application which contains your preferred normal fonts at switch-
on, and a directory called spare (for example) which contains the set of
fonts that you wish to switch in and out. Then in your !Boot file, the
FontInstall command becomes
6.9
FontInstall <Obey$Dir>.fonts.
6.9
and the IF...THEN statement in the !Run file has <Obey$Dir>.spare. in
place of the two occurrences of <Obey$Dir>. .
6.9
Miscellaneous Comments
6.9
FontEd
6.9
When Tom Rank spoke to an Acorn representative he was told that they
were not contemplating the release of FontEd and that it is NOT public
domain. An early version got there by mistake via SID and it really
ought not, he said, to be distributed by PD libraries.
6.9
The Acorn representative acknowledged that Acorn were unlikely to do
anything about this situation, as it was “only an early version and
won’t work on RISC OS 3”. According to Tom this is partially true; a
friend of his has had problems with version 0.28, but at least some of
the time it runs on an A5000. However, he observes that it is a pity
that Acorn actually seem to be discouraging anyone else from working on
fonts. (There is a version of FontEd on Careware 7, released in May
1990, but it has been updated since then. Acorn haven’t objected,
presumably because of the charitable aspect. Ed.)
6.9
Screen blanking during floppy disc access
6.9
Stuart Bell writes: “Many things are very much more elegant in RISC OS 3
but with one notable exception. I couldn’t believe my eyes when I saw
what happens to the screen during floppy disc accessing. Whether my ARM3
cache is dis- or enabled, in mode 24 the screen either flashes or is
blanked when loading, saving or formatting. If RISC OS 2 didn’t need to
do this, why is it allowed now? (It did, in fact, happen in mode 21
under RISC OS 2 – HE.) Some re-coding is needed, even if the resultant
code is a little slower. More significantly, who was the Acorn person in
authority who thought that it is acceptable in 1993 to produce a
computer which professes to be a ‘world-beater’ which, when using floppy
discs, looks like the Sinclair Spectrum (or was it the ZX81?) reading
from cassette tape circa 1983? This just looks so tacky. I know that
serious use demands a hard disc, but first impressions count for a lot,
and a lot of demonstrations will be on floppy-based A3010’s. This is not
good enough, Acorn.”
6.9
Basic program compaction errors
6.9
R. W. Darlington writes: “I am writing on a matter of utmost urgency.
6.9
“There seems to be a growing urge amongst programmers to use a program
compactor on Basic RunImage files. Unfortunately, I know of no Basic
compactor program that does not exhibit a certain disastrous fault.
6.9
“If, for instance, a Basic program contains any floating point or pseudo
variables followed by a Basic keyword or statement, then when it is
compacted and the space between the variable and the keyword is removed,
that line will generate an error.
6.9
“For instance, if the program contained the line
6.9
10 IF TRUE THEN VDU7
6.9
which when compacted becomes
6.9
10IFTRUETHENVDU7
6.9
the lack of space between TRUE and THEN will generate an error.
6.9
“The same is true for any line containing a floating point variable:
6.9
10 ha=TRUE
6.9
20 IF    ha    THEN    VDU7
6.9
will crash when compacted because of a lack of space between ha and
THEN.
6.9
“I have come across numerous examples of RunImage files which have these
errors in them. One of them is Acorn’s new RISC OS 3 !SciCalc. Others
are new versions of Translator and Creator.
6.9
“Please can you draw Archive readers’ attention to these problems as
soon as possible? Or, if any reader knows of a Basic compactor that does
not produce these errors, please can they let us know where it may be
obtained?”
6.9
Matters Arising
6.9
Setting Caps Lock (*FX 202)
6.9
Unfortunately, the advice I gave last month about using *FX 202 may have
been a bit confusing because, after explaining that Peter Young wanted a
command to turn Caps Lock on, I then described how to turn it off! The
command to turn Caps Lock on, leaving the other settings unchanged, is
*FX 202,0,239.
6.9
The shortened form *FX 202,32 should turn Caps Lock on, but may have
side effects.
6.9
Note also that, to make sure that the LED indicators are correctly
updated after a *FX 202 command, it is a good idea to follow it with the
simple command *FX 118.
6.9
The bits of the keyboard status byte have the following meanings:
6.9
    bit    value    meaning
6.9
    0 (1)    0    Alt off
6.9
        1    Alt on
6.9
    1 (2)    0    Scroll Lock off
6.9
        1    Scroll Lock on
6.9
    2 (4)    0    Num Lock on
6.9
        1    Num Lock off
6.9
    3 (8)    0    Shift off
6.9
        1    Shift on
6.9
    4 (16)    0    Caps Lock on
6.9
        1    Caps Lock off
6.9
    5 (32)        (“Normally set”)
6.9
    6 (64)    0    Ctrl off
6.9
        1    Ctrl on
6.9
    7 (128)    0    Shift enable off
6.9
        1    Shift enable on
6.9
(This information is taken from the RISC­OS 2 Programmers’ Reference
Manual.)
6.9
Boot-up options (Archive 6.8 p26)
6.9
Peter Young has suggested that it might be better to use INKEY$ with a
delay of, say, 5 to 10 seconds instead of GET$ so that, if nobody
presses a key, the computer will continue with the boot file after a
short delay.
6.9
Unexpected messages from IDE filing system (Archive 6.8 p24)
6.9
Peter Young (who has a Beebug IDE interface) has confirmed that, even
after creating space for a directory cache, he still gets unexpected
messages unless he uses full pathnames. For instance, if he starts with
the commands
6.9
*DIR IDEFS::IDEDisc4.$.BASIC.B
6.9
*CAT
6.9
he gets a catalogue of the subdirectory ‘B’ as expected. However, if he
uses the abbreviated pathname instead:
6.9
*DIR $.BASIC.B
6.9
*CAT
6.9
he gets a catalogue of the directory ‘BASIC’ even though the heading of
the directory listing correctly says that the CSD is IDEFS::IDEDisc4.$.
BASIC.B. The computer’s confusion is further demonstrated by the fact
that if he then enters:
6.9
*DIR ^
6.9
*CAT
6.9
it replies Directory ‘@’ not found.
6.9
‘Interrupt’ module (Archive 6.6 p55)
6.9
In March, we reported M. P. Sawle’s request for help in getting the
Interrupt module (Archive Shareware disc 6) working under RISC OS 3, but
since then neither he nor we have heard anything. Does no-one else use
this module? Can its author help?
6.9

Using RISC OS 3
6.10
Program points
6.10
•    QuicKey – Last month (Archive 6.9 p 61), we reported a problem with
this program. Ian Copestake has written to say that this problem was
cured by an update released in April 1992 – over a year ago! – and was
caused by a bug in RISC OS 3. He says that customers can always obtain
minor updates like this by returning their original disc plus £1 worth
of postage stamps (or a reasonable overseas equivalent). He also asks if
contributors can check that they have up-to-date versions before
reporting problems.
6.10
•    Fourth Dimension games – A. H. Hunt had a problem with Black Angel
(“one of the casket segments was missing at the Planet Valve (23%)”) but
this was rectified by 4D immediately. He also says that Break147 and
Saloon Cars run (for him) without any problems.
6.10
However, Arthur Taylor found that the hint by D. S. Allen in Archive 6.9
p62 on Break147 and SuperPool didn’t work for him. Perhaps
significantly, the relevant line of the Basic program was line 100 not
line 90 in his copy, and also his program wasn’t compressed but was
heavily protected. (I wonder whether the version of Break147 which A. H.
Hunt has would work on his machine, and whether Fourth Dimension can
supply it?)
6.10
•    Rotor − This game from Arcana does not work on RISC OS 3 and Arcana
have, apparently, gone bust. Fortunately, Superior Software have bought
out the rights to the program and sell it as part of their “Play it
Again, Sam 1” compilation. Apparently, if you send your Arcana version
of Rotor to Superior with a cheque for £10, they will send you a copy of
Play it Again, Sam 1.  John Waddell, Renfrewshire.
6.10
•    Recommended programs (which do work) – Arthur Taylor strongly
recommends Simon Burrows’ FontsPlus: he has created four fonts
directories and with FontsPlus on the iconbar he can choose to have all
or none of them active (if none, the ROM fonts are still available); if
he doesn’t load FontsPlus, then they are all available “... simple as
that!”
6.10
He says that other PD or shareware programs that seem to work fine with
RISC OS 3.10 include: DiskBase v0.65, DiskCat v1.50 (disc cataloguers),
FuncKeys v2.02 and KeyWindow v1.40 (keystrips and keys from the desktop)
and a super neat new mode changer from Dave Thomas: !Mode v1.03 – it can
store up to 8 modes in a small desktop window and seems very well
behaved.
6.10
•    Watford scanner − In Archive 6.8 p27, we reported problems with the
Watford Handscanner (1990, issue 2). Apparently, there is an upgrade to
the software (version 1.1) available from Watford.  D Webb, Cumbernauld.
6.10
Disc drive problems
6.10
•    360Kb PC discs − Richard Fallas writes: “While trying to transfer data
to the hard drive of an old PC recently, I discovered that there are no
hard and fast rules about compatibility with 40 track discs. Previously,
I had used my switchable 40/80 track 5¼“ drive (with Beebug Interface)
under the emulator, having set up a suitable device in the config.sys
file. This no longer works under RISC OS 3.10, so I tried to transfer
directly.
6.10
“The disc format was correctly identified and the files written (with
switch set to 80 track) without hitch, but the PC didn’t want to read
them (using copy *.*), although it would give me the directory listing.
6.10
“Being stubborn, I tried again, saving and copying file by file. This
worked... sometimes! The longer the file the greater the chance of a
glitch, but eventually I got all the files across. This worked on files
up to 75Kb in size, so persevere!”
6.10
•    IDE drives – Arthur Taylor advises that you should not assume that the
IDE support built into RISC OS 3 will work with an IDE drive that has
been added to a pre-A5000 computer. He found with his IDE drive supplied
by DT that he had to set the number of IDE discs to zero in !Configure
and then run DT’s setup, after which he had no problems.
6.10
•    ST506 drive fitted to an A3000 – Tony Flaherty was interested to read
of the hardware problem experienced by Colin Wood (Archive 6.9 p61) when
fitting a ARM3 to his A410, since he has recently experienced a similar
problem with his A3000. He writes:
6.10
“I have a basic ARM2 machine with RISC OS 3.1 and an Orion 2 Mb upgrade.
I recently added an external 20 Mb ST506 hard disc using an ex-PC MFM
drive and an Acorn external controller. The drive and a suitable power
supply/fan sit in a home-made expansion box.  I suffer the usual A3000
problems with regard to the oxidation of the memory connectors.  The
hard disc also tends to generate defects if the system is powered down
in an abnormal way (i.e. not <ctrl-shift-f12>).
6.10
“My problem was symptomatically the same as Colin’s in that, initially,
the system froze completely while I was editing a First Word Plus
document eventually forcing me to press <reset>. The screen flashed red
then blue with no further action − a power down produced the same
results, I tried this several times to no avail. Finally in desperation,
I opened the box but nothing looked out of place (no charring or loose
leads).  I powered up the machine and this time all was well, however,
when I verified the disc 20 or defects were reported, when these were
added to the disc’s defect list, using *DEFECT, some of the files within
the first Word Plus application directory where corrupt and this had to
be restored from a backup.”
6.10
Tony wonders, therefore, whether Colin Wood’s assumption that his
problem related directly to the ARM3 upgrade was correct.  Could the
problem lie with RISC OS 3.1?  Does anyone have any views on this?
6.10
Other matters arising
6.10
•    Cache control − In response to the comments about difficulty on
RISC OS 3 with hares, tortoises and ARM3s, David Percival has sent in an
application that controls it properly. It evens notices if you have
typed in a cache-on or cache-off command or have executed one within a
program and it still displays the correct icon on the iconbar!  David
Percival, Swanley.
6.10
•    Redraw problems (Archive 6.8 p29) − A problem of incorrect screen re-
drawing in !Edit was mentioned. Tim Nicholson has had the same problem
with Deskedit! He says it seems to be a particular problem with
unformatted text that only has carriage returns/line feeds as paragraph
breaks and nowhere else. Since both pieces of software have the same
problem, he wonders if the bug is in the SharedCLib? Has anyone else had
the same problem?
6.10
•    Printing after Artworks (Archive 6.5 p31 and 6.8 p24) – Roger King
reported a problem with spurious fine horizontal lines about 4mm long
scattered throughout printouts when printing from another application
after printing from Artworks. Steve Hutchinson found that putting his
LaserDirect podule in a different slot (as recommended by Computer
Concepts) solved this problem for him. Unfortunately, this didn’t work
for Roger. However, he has now found a cure: after printing from
Artworks, select “Queue control” from the printer icon and then select
“Flush buffer”. He observes that it would seem that Artworks is leaving
something “nasty” in the printer buffer, although a CC representative
with whom Roger discussed the problem suggested that Acorn’s printer
drivers might be at fault.
6.10
•    Photocopying via LaserDirect (Archive 6.8 p24) – Tim Powys-Lybbe has
got version 2.10 of the LaserDirect printer driver which works quite
satisfactorily and he has no more problems “photocopying” from Scanlight
Plus.
6.10
He adds cryptically that “some things may not print totally as one would
expect in Paint or Scanlight Plus, particularly their positioning on the
page, but putting them into Draw or Impression totally solves the
problem.”  A
6.10
Hints and Tips
6.10
•    Altering Impression frames − Here is another one of those hints that
everyone but me probably knew about all the time: If you want to change
the attributes of several different frames, (like making them all
transparent) all you do is click on the first frame, call up Alter frame
(with <ctrl-f10>), make the changes and then click OK with <adjust>
instead of <select>. This keeps the Alter frame dialogue box on screen.
Then click <select> on the next frame to be changed and make the change
in the Alter frame dialogue box which is still there on screen. This
speeds things up no end.  Jochen Konietzko, Köln, Germany.
6.10
•    Background printing − Various comments in Archive have suggested that
you need the Ace PROdriver in order to get background printing. However,
Keith Sloan’s PD program !Spooler (on Shareware 24) works with a
Panasonic 1124 + Impression II on RISC OS 3.11 − cheaper than using a
PROdriver!  Chris Dawson, Derby.
6.10
•    Disc storage boxes − Looking for new disc storage boxes? Look no
further than your local Supermarket or D.I.Y. Superstore! I have
recently purchased the Module 2000 Box and Lid (Ref. no’s. Unit 2 and
Lid 2) manufactured by Addis, the plastic kitchen accessory
manufacturers. The size of the box is 450×190×127.5mm and comes in
various colours. The combined cost of the box and lid recently was £4.78
and it holds approximately 144 discs. (I have arranged them in four
rows, each holding 36 discs). What is more, it is made in Britain, like
the Acorn Archimedes, and better in style and strength. Give me Addis
any time!  Timothy Brown, Kent.
6.10
•    Floptical drives − I offer the following as ‘Hints’ arising out of the
fairly extended process of getting my floptical drives to work (with
much help from Morley), in the hope that they will be useful to other
new owners.
6.10
The drive will not work at all with older ROMs fitted to the SCSI
interface board − you need version 1.12 at least. If you already have a
SCSI hard disc (or other SCSI peripheral) there should be no termination
resistors in the floptical drive. They are easily removed but you have
to remove the drive from its case first − they are close to the ribbon-
cable connector on the underside of the drive.
6.10
The drive is recognized as two devices. The 21Mb disc counts as a hard
disc and appears as SCSIDrive5 in my case (as I already had a SCSI
drive), while the facility for reading and writing to ADFS discs counts
as a floppy drive. I could not get the floppy device to be recognized at
all using the SCSI_conf program on Morley’s support disc, but had to do
it manually, using: *configure scsifslink 0 2 (followed by a reset).
This gives a ‘floppy’ icon called SCSIDrive0. (My drive arrived set up
as SCSI device 2 − you would need to modify the configure command if it
was set up to be another number.) There is no conflict between the SCSI
and ADFS floppy drives, although they are both called 0, and anyway the
computer will not allow you to call the SCSI floppy drive 1 if there is
no SCSI floppy drive 0.
6.10
The writing under the SCSI drive icons changes from the drive name to
the disc name when a disc is mounted. This is potentially confusing
until you know about it, because the default name of the 21Mb disc is
the number of the drive. I got into even worse confusion because I had
tested the Floptical on its own, so the disc became labelled “disc4”.
When I restored the real SCSI hard disc as well as the Floptical, I got
an error message saying there was a conflict of disc names.
6.10
By the way, since the floptical drive will not format discs, does anyone
know of a source of high-density floppies formatted ADFS 1.6Mb? Philip
Draper, Borehamwood.
6.10
In view of this problem with flopticals being unable to format 1.6Mb
discs, we are offering pre-formatted discs at £17 for 10. Ed.
6.10
•    Pocket Book battery connections − The middle battery connector on the
Pocket Book/Psion 3 does not allow the use of AA batteries with indented
positive and/or negative terminals (e.g. Ever Ready’s rechargeable
cells.)  Paul Bisonnette, Otterfing, Germany.
6.10
•    SigmaSheet does not return to the desktop neatly. If programs such as
Ovation are running, their Alt-key combinations will be lost when
SigmaSheet returns. To solve this, amend the SIGSHEET program as
follows. Look at the end and find DEFPROCexit and change it to read:
6.10
DEFPROCexit
6.10
*FX 4
6.10
ON:*FX 225,1
6.10
*RMReInit InternationalKeyboard
6.10
IF INKEY−2 : END : ELSE *Quit
6.10
John Waddell, Renfrewshire.  A

Hints and Tips
6.11
•    Amstrad CPC monitor link − Let me describe how to use the old monitor
left over from your Amstrad CPC for emergency use with your Archimedes.
6.11
You need a six-pin, 270-degree DIN female plug, like the one in your
CPC, a normal D-sub nine-pin male, to plug into your RGB socket on the
computer and some soldering skill. (If you use an A5000 or later, you
need a 15-pin VGA-compatible plug, instead.)
6.11
To make the setup even more user-friendly, you should invest in a two-
way, double switch, to cater for both mono and colour monitors.
6.11
The DIN plug has six pins, with the centre pin being mono-in. When seen
from the rear with the soldering lugs towards you, (as in the diagram
opposite) the pin for the red is top left (1), then comes green (2),
blue (3), sync (4) and signal earth (5).
6.11
On the 9-pin D-sub, pin number 1 is red, 2 is green and 3 is blue. Pin
number 4 is sync and 7 signal earth. The rest are unused. The
arrangement on the new machines is slightly different, as there are more
pins, but it is essentially the same!
6.11
Basically you connect 1 to 1, 2 to 2 and 3 to 3, but the addition of a
switch makes it much more useful. Before starting, switch on the
computer while pressing <0>, to change to monitortype 0 for 50Hz
monitors and TVs.
6.11
The switch on the cable is essential under two circumstances: You might
also want to use your TV as a colour monitor. Then you have to use the
Amstrad TV MP-2 modulator (still available). That does not use the
central monochrome pin, but the normal three RGB pins. You could make
two converters of course, but you might forget which is which. Instead,
you just flip the switch if the picture looks odd! Those with TVs that
accept RGB through the SCART connector should use those, of course!
A3010 owners should first try connecting to a TV, but the green CPC
monitor should be much sharper, if not having a higher resolution!
6.11
Sometimes, when running some games like MiG-29M, you might want to
switch back and forth as the headup display is almost invisible against
the sky if you have all colours connected to the middle pin. By
momentarily switching to the colour setting, everything yellow will
become easily visible, as it becomes black (or more correctly, a black-
ish green, if you use the green monochrome monitor)!
6.11
Turbo Challenge II becomes a real challenge in green. Chopper Force,
Bughunter II and Taipei work fine in green. Impression, Vector and
ArtWorks are essentially monochrome, if you don’t add colour yourself,
so they are not affected!  Tord Eriksson
6.11
•    BASIC64 application − I have, for a long time, toyed with the idea of
testing BASIC64, by running the PCW Benchmarks with it. The normal way
would be to RMLoad the BASIC64 module and then type *BASIC64 − not very
user-friendly, as you have to type in the paths each time.
6.11
Remembering David Pilling’s advice on the uses of the OSCLI command
(very odd to me, brought up on Spectrums and mainframes!) I wrote the
following Basic lines:
6.11
*DIR SCSI::Conner170M
6.11
*DIR OldRubbish.BASIC.BenchMARK
6.11
*CAT
6.11
INPUT “Which file? ”;a$
6.11
OSCLI “*BASIC64 -quit ” + a$
6.11
The *CAT shows me which files are available and OSCLI “*BASIC64 -quit”
loads BASIC64 runs the program a$ and then quits. Your version will have
to include your drives and directories, of course! This programming
fragment, I called !RunImage.
6.11
Put this in the directory you have made called !Run64.
6.11
To make this an application, steal some !Sprites from a program you
like, rename the individual icon sprites to !Run64 and sm!Run64.
6.11
The next step is to make a copy of BASIC64 and drop that into the
directory. Now for the !Run file:
6.11
Iconsprites <Obey$Dir>.!Sprites
6.11
RUN <Obey$Dir>.BASIC64 -QUIT <Obey$Dir>.!RunImage
6.11
Alternatively, you could run the copy of BASIC64 that you might have in
your system’s modules directory.
6.11
It seems idiotic to run BASIC64, in order to run BASIC64, to run the
program you have selected, but it works perfectly.
6.11
I tried using the more logical *RMLoad BASIC64, but then you fall into
the editor and we do not want that!
6.11
Finally, you need a !Boot file:
6.11
Iconsprites <obey$Dir>.!Sprites
6.11
(A sample Basic64 application is on this month’s program disc.)  Tord
Eriksson
6.11
•    ColourCard and VIDC software − After I had installed my ColourCard and
(thanks to a hint in Archive) realised that I still would have to use a
VIDC module (Careware 18) to make my monitor display modes 12 and 15, a
problem arose. I had configured my machine to one of those wonderfully
large screen modes (1152×848 logical pixels), and then when I found the
time to play a mode 15 game, (Tower of Babel, which is completely mouse-
controlled), there was no pointer!
6.11
This is what Nigel Star of Wild Vision told me: In a ColourCard mode,
the VIDC pointer is disabled and some games don’t re-enable it so all
you have to do is switch to mode 15 first, and only then start the
game.  Jochen Konietzko, Köln.
6.11
•    Combatting fiddlers − A solution for teachers who suffer from children
‘fiddling’ with the palette, task windows, applications accessing the
hard disc and destroying files might be to use the RMKill command in the
!Boot file. Any or all of the following lines can be added to the !Boot
file.
6.11
RMKill ResourseFiler
6.11
RMKill ADFSFiler
6.11
RMKill PaletteUtil
6.11
RMKill Taskmanager
6.11
This will leave the iconbar empty. Pressing <reset> will give error
messages. If you want a particular program to run, this could be done
from the !Boot file or from an application launcher such as !Menon or
!Two. Directories can be opened in the same way. Switching on whilst
holding down <shift> avoids running the !Boot file and returns icons to
the iconbar.  Paul Harrigan, Hong Kong.
6.11
•    DragASprite module − After seeing the tip on the dragging of solid
sprites from a previous Archive, I decided to explore the DragASprite
module for myself. I have not worked on it exhaustively but I have come
across the following which users could put in their Basic programs.
6.11
The DragASprite module supports two SWIs but the more useful of these
two is the SWI DragASprite_Start. It requires the following parameters:
6.11
R0    control value
6.11
R1    1 (this seems to be the only value that works)
6.11
R2    pointer to sprite name
6.11
R3    points to start coordinates of the drag box
6.11
The control value consists of bits which affect the drag. The only ones
I have worked out are:
6.11
bit 6    allow dragging outside graphics window
6.11
bit 7    shadow the sprite
6.11
Bits 0 and 4 seem to allow only the value 5, as far as I can work out.
6.11
I have found that the two most useful values for the control value in R0
are:
6.11
&85    non-shadowed drag outside the graphics window
6.11
&C5    shadowed drag outside the graphics window
6.11
R2 points to the first character of a string which is the name of the
sprite to be used for the drag, for example “file_fec”.
6.11
R3 points to four 32-bit coordinates which describe the start and size
of the drag box (which is the same size as the sprite).
6.11
This is about all I have been able to work out in the time I have spent
on it, but it is possible to modify Basic programs using the Save As
window to use DragASprite_Start instead of Wimp_ DragBox.
6.11
For example, in FormEd 1.24, modify line 4810, which normally reads:
6.11
4810 SYS “Wimp_DragBox”,,q%
6.11
to
6.11
4810 SYS “DragASprite_Start”,&C5 ,1,“File_fec”,q%+8
6.11
I’m sure a similar fix could be used for other programs to give solid
sprite dragging.  Robin Terry, Newcastle upon Tyne.
6.11
(Those who can afford £100 for a RISC OS 3 PRM will find more details on
page 3-299. Ed.)
6.11
•    Free space on hard discs − In Hints and Tips, Archive 5.12, Steve
Drain asked if the same problem of allocating 28Kb to each new directory
applies to IDE drives. I checked this for my A5000 40Mb drive and found
that 7Kb is allocated per directory which seems a reasonable compromise.
This is lucky as there seems to be no way to alter the default format
using HForm.
6.11
N.B. Count only reports the size of the data, not the actual space used
on the disc. This can only be found by using *FREE from the command line
before and after creating a directory/file and subtracting the values.
6.11
The structure of !Fonts is where most directories occur. In my case,
there are 111 directories requiring 777Kb of space initially, but since
each contains over 7Kb of data for IntMetrics and Outlines, the space is
not wasted. However, if 28Kb were allocated, as with Steve’s make of
SCSI drive, then 111 × 28 = 3108Kb would be allocated. In my case, only
2625Kb would be used as font data, losing 483Kb on fonts alone! When
Computer Concepts released Compression, I thought that this would solve
space problems, but I was told by CC that it retains the directory
structure (I assume with the same overheads per directory) and is not
very effective at reducing the disc space taken up by !Fonts.
6.11
I have now installed SparkFS from David Pilling. Although each compacted
Spark file can contain lots of directories when opened, to the IDE
filing system it is only one file and only allocates one lot of 7Kb (or
28Kb for Steve). So it is not sensible to have a Spark file containing
less than 7Kb. (The Count facility is misleading as it only gives the
size of compressed data not the space allocated by the IDE drive.) There
is a speed penalty for using a Spark file but there is also a Spark
directory option which is a lot faster according to the booklet although
it seems to have a much larger disc space requirement of about 14Kb
which does not all seem to be recovered.
6.11
So the conclusions seem to be:
6.11
1)    If possible, reformat with the smallest file allocation.
6.11
2)    Only create a new directory if you are going to store an amount of
data comparable with the directory space (7Kb for my IDE drive, 28Kb for
Steve’s original SCSI).
6.11
3)    Be aware of the overheads due to any compression system you use,
checking by using *FREE at the command line (particularly if it retains
directories.)
6.11
4)    If using SparkFS on an A5000, ensure any Spark file will contain
at least 7Kb of data.
6.11
5)    If space is tight, remember that a Spark directory will have space
overheads although it is faster.
6.11
It would be interesting to hear the findings from other people using
other compression systems, types of drive and any information on
reformatting A5000 IDE drives with different file allocation sizes.
Alan Dawes, Ilford
6.11
•    Help_ with menus in C − In Archive 6.10 p8, Richard Simpson asks about
using help_ with menus. I enclose a program (on the monthly program
disc) to demonstrate a solution.
6.11
The program offers help via !Help in the following situations:
6.11
  •    The main iconbar icon.
6.11
  •    Menu (with sub-menu) from from the iconbar.
6.11
  •    Dialogue box from iconbar, including special processing for
version details.
6.11
  •    A simple window.
6.11
  •    Menu from the simple window.
6.11
The program doesn’t actually do anything and is quite long, so I won’t
attempt to explain its operation here. There are a number of comments in
the code which should, hopefully, be sufficient for a C programmer using
help_.
6.11
The short-hand strings supported by !Help are (this is from !Help’s
message file), note the spaces:-
6.11
\S  ’Click SELECT to ’
6.11
\A  ’Click ADJUST to ’
6.11
\G  ’This option is greyed out because ’
6.11
\D  ’Drag SELECT to ’
6.11
\w  ’window’
6.11
\a  ’ADJUST’
6.11
\R  ’Move the pointer right to ’
6.11
\T  ’This is the ’
6.11
\W  ’This window is ’
6.11
\d  ’Drag ADJUST to ’
6.11
\s  ’SELECT’
6.11
Please bear in mind that the documentation in Release 4 is sketchy and
this may not be Acorn’s preferred solution, but it works. Hope this is
of some use.  Joe Wood, Goring-by-Sea.
6.11
•    Impression master pages − To create a document with the same master
page as an earlier document in Impression, load the first document, save
a new copy with a new name, create a new chapter with the following
choices: Insert after current chapter / Restart page at 1 / Copy master
page and delete all of the chapters except the new one.  George Foot,
Oxted.
6.11
•    Numeric resets − I have a 420/1 and, because of lack of space on the
table where the computer lives, the keyboard has to be kept upright
behind the computer when it is not in use. One day, I turned the
computer on and was greeted by a display moving rapidly upwards and
slowly sideways. Adjusting the monitor had no effect. It turned out that
one of the numeric keys was being held down by the corner of the monitor
and the monitor type had been reconfigured as I switched on. So, if you
keep your keyboard upright (yes, I know it’s bad practice, but where
else can I keep it?) and your display goes crazy, try a <numeric> reset
before panicking!  P Young, Cheltenham
6.11
•    Pocket Book and BJ10ex problems − It appears that the Pocket Book’s
built in Canon BJ10e driver does not work correctly with the Canon
BJ10ex. We have a printer driver (supplied to us from Psion via Richard
Collinson) which works. If you have this printer, send us a formatted
SSD or a disc and your problems will be over. Psion can also supply
other drivers on SSD if necessary.  Simon Moy, Archive.
6.11
• SigmaSheet revisited − John Waddell’s SigmaSheet fix in last month’s
Hints & Tips requires one more line within DEFPROCexit: *Country UK.
6.11
•    Solid icons − If the ‘drag solid file icons’ option is set in RISC OS
3, it causes applications like DrawPlus to reject the file as being of
an incompatible file format. This prevents files being dragged between
applications in the normal way.  Mike McNamara, Dunstable.  A
6.11
Using RISC OS 3.10
6.11
Hugh Eagle
6.11
Neil Dickson asks a jolly good question: have I or Archive ever had any
response from Acorn indicating that they are paying attention to this
column? Well, I haven’t (yet) and I think Paul would have told me if he
had. Surely, it’s inconceivable that Acorn are not interested in the
experiences that have been recounted, the suggestions offered and the
views expressed here. Isn’t it? Perhaps we’ll have some comment for next
month. Let’s wait and see ...
6.11
RISC OS 4 Wish List (continued)
6.11
•    Alarm – (a) An option to display the digital clock not on the iconbar
but at the top right corner of the screen in front of all other windows.
(b) An option to set an alarm that goes off after every n (settable)
minutes − to remind you just how much time you are spending on the
machine!!?? (Rex Palmer)
6.11
•    Shutdown – The <shift-ctrl-F12> routine should dismount any floppy
discs that may be present, then remind the user that they are still
there. (Peter Young)
6.11
All the following are from Neil Dickson:
6.11
•    Paint – (a) Allow local fills with the “mask” colour. (b) Use outline
fonts.
6.11
•    Edit – (a) Use multiple fonts (and maybe accept graphic files: i.e.
have a simple wordprocessing option). (b) Update the font list each time
a new font directory is loaded (like Draw and Impression do.)
6.11
•    Printers – (a) Put as much as possible of the code into ROM, to save
loading time. (b) A quicker method of “toggling” within !Printers when
deciding which are active and which are inactive.
6.11
•    Calc – Make it better, but less complicated than SciCalc. (E.g.
include memory buttons and a percentage button, and allow
multiplication, division, etc, by a constant.)
6.11
•    Draw – (a) Optional detachable toolbox. (b) An option for redefining
the toolbox (e.g. to include zoom and edit tools.) (c) Patterned fills
(hatching, stripes, etc.)
6.11
•    Tidy windows – Include a “Tidy windows” option on the desktop.
6.11
•    Filer multitasking – An option to make all filer actions (load and
save, disc mounting, etc) multitask.
6.11
•    Replacement of system font – Optional replacement of the system font
by an outline font on the desktop (like the Germans seem to have already
in RISC OS 3.12)!
6.11
•    RMFaster – An option to RMFaster various modules, such as
SharedCLibrary, from the Task Manager (rather than having to type in the
command at the command line).
6.11
•    Virtual memory – For hard disc users.
6.11
•    Built in sounds – A greater variety.
6.11
•    Simpler mode changing – A less fiddly method.
6.11
•    Keyboard emulation of the mouse – e.g. using <Alt> and the arrow keys.
6.11
•    Auto-scrolling of font menus – Make the list of fonts scroll
automatically when the pointer reaches the bottom.
6.11
•    Compression of sprite files – Either in addition to, or instead of,
compression of entire discs, arrange for sprite files to be compressed
always.
6.11
•    Font management – Put a Font Manager on the icon bar showing a list of
available directories which you can select and deselect by clicking.
(Sounds like !FMaster from Risc User.)
6.11
Miscellaneous
6.11
Superfluous icons in template definitions
6.11
Mick Day has found numerous examples of template definitions in standard
RISC OS 3 applications which contain extra unused icons. The “save”
window in Alarm is just one example: this has four superfluous icons.
(To see for yourself: find the template file inside Resources:$.
Resources, load it into FormEd, use the menu to assign vertical and
horizontal scroll bars and a size icon, then scroll around the window.)
He has also noted inconsistent naming of templates and numbering of
icons. If you use existing windows borrowed from existing applications,
beware that the templates may require more memory than you might expect
because of the space taken by the extra icon definitions and pay close
attention to the icon numbering. As Tim Nicholson comments: it is a
great pity that Acorn did not supply a suitable “clean” set of templates
to work from.
6.11
“OS_Find” and image filing systems
6.11
In one of my programs, I need to be able to check that a specified
filing system object (which might be either a file or directory) is
actually present and, in particular, is not on a floppy disc that has
been removed. In RISC OS 2, I found that the following code seemed
fairly foolproof:
6.11
ON ERROR LOCAL:PROCReport(“Can’t find ”+Object$):ENDPROC
6.11
SYS “OS_Find”,&40,Object$ TO FileHandle%
6.11
IF FileHandle%<>0 THEN SYS “XOS_Find”,0,FileHandle%
6.11
RESTORE ERROR
6.11
This relied on the fact that the SWI call “OS_Find”,&40 (which is the
general routine for opening a file) should “open” Object$ whether it is
a directory or a file. However, with the arrival of “image” filing
systems such as DOSFS and SparkFS, I found that this wouldn’t work,
because an attempt to “open” a directory inside an image filing system
(e.g. a directory inside a SparkFS archive) resulted in an error
(contrary to the rules set out in the Programmer’s Reference Manual).
6.11
David Pilling says “there are two entry points to SparkFS, one via the
OS image filing system, the other via a traditional FS entry. The code
in these after the entry point is the same. So it looks like it is the
OS that is faulting your operation and not SparkFS.”
6.11
I have now rewritten my code so that it checks to see whether Object$ is
a directory and, if it is, it looks inside it for a file and tries to
open that. This seems to work with image filing systems as well as
others, but is a bit cumbersome to say the least (and as David Pilling
observes is still open to being fooled by cacheing). Does anyone know of
a simple and genuinely foolproof way of ensuring that a filing system
object is present?
6.11
Applications that refuse to Shutdown
6.11
Keith Gardner has experienced problems with applications that interfere
with the Shutdown process and cause the computer to hang. He refers to
advice in the June issue of Risc User, according to which the Shutdown
routine in RISC OS 3 (which can be selected from the Acorn icon or by
pressing <ctrl-shift-F12>) cannot cope if more than one Run file is open
at the time of Shutdown. Run files are left open while applications are
running if there are any lines after the line that calls RunImage. One
way to avoid the problem is to remove surplus blank lines or lines
containing remarks (prefaced with a | ). However, applications sometimes
contain tidying up commands such as RMKill commands to get rid of
modules: in these cases, it may be necessary to Quit one or more
applications before selecting Shutdown.
6.11
Matters arising
6.11
Testing a printer’s on-line status
6.11
(Archive 6.9 p12) Mick Day found that the printer test routine published
in the June edition of Archive lied! It told him that the printer was
not on line whether it was or not, and the printer beeped if it was on
line. Tim Nicholson says that the routine (as previously published) can
fail for two reasons: Firstly, it assumes a printer buffer of 1024
bytes, but the actual size depends upon the setting of *Configure
PrinterBufferSize, which may be different from the default of 1Kb.
Secondly, the first line of the program redirects the whole output
stream to the printer, thus if the program is run from the desktop, the
output which normally goes to the Task window that the desktop
generates, goes to the printer instead as well as the test byte of
CHR$(0). Typically this stream will contain several control codes to
open the task window display and the Task window title, e.g. RUN
ADFS::DiscName.$.PrintTest. Although this stream is sent before running
the program, the FX 3 call at the top of the program is executed in time
to redirect this part way through and upset the printer.
6.11
Tim suggests the following variant of the routine:
6.11
*FX21,3
6.11
empty%=ADVAL(-4)
6.11
PRINT “Printer is ”;
6.11
*FX3,10
6.11
PRINTCHR$(0);
6.11
a%=INKEY(20)
6.11
a%=ADVAL(-4)
6.11
*FX3,0
6.11
*FX21,3
6.11
IF a%<>empty% THEN
6.11
PRINT “not on line”
6.11
ELSE PRINT “on line”
6.11
ENDIF
6.11
END
6.11
This deals with the first problem by comparing the buffer length
immediately after flushing with the length after inserting the test byte
and pausing to allow throughput, and with the second by including a
PRINT statement before redirecting the output stream.
6.11
Tim notes that the original program would run correctly (subject to the
buffer size being correct) from the command line, i.e. after pressing
<f12>, and CHAINing or LOADing and RUNning.
6.11
Tim does not understand the supposed warning from Acorn about the use of
ADVAL(−4) (see Archive 6.7 p41) since it is a use specified in the Basic
manual. He can only assume it is a confusion with the use of positive
ADVAL parameters, which are only meaningful if a podule with an analogue
input which mimics that of the original BBC Micro is fitted. He has
experienced no difficulties using negative ADVAL parameters to inspect
the condition of buffers and is not aware of any differing action under
RISC OS 3.1x as opposed to 2.0 with this facility.
6.11
Basic program compaction
6.11
(Archive 6.9 p65) Rex Palmer says that in his experience compacted
programs only cause problems when you try to alter them. The problem
seems to occur with both Edit and BasicEdit. If a compacted program is
loaded into either of these and any alterations made, then when the
program is re-saved, the line that has been altered will generate the
error even if the resulting error is not part of the line that was
altered!
6.11
Altering the keyboard status
6.11
(Archive 6.8 p27 and 6.9 p65) Rex Palmer has written to explain the
purpose of bit 7 (the “Shift enable” bit) of the keyboard status byte.
When this bit is set, the <Shift> key reverses the action of the
<CapsLock> key. Thus, if you hold down <Shift> while pressing <CapsLock>
then release both keys, you put the computer in a state where pressing
the <Shift> key with a letter key gives you lower case letters. The
computer can be configured to start up in this state. Alternatively, the
command *FX202,128,111 will achieve the same effect ... since 111 = 255
− 128 − 16 the command sets bit 7 (Shift Enable on) and unsets bit 4
(Caps Lock on).
6.11
To recap, the most useful commands are likely to be:
6.11
*FX202,0,239 to set Caps Lock on
6.11
*FX202,16,239 to set Caps Lock off
6.11
*FX202,128,111 to select ShiftCaps
6.11
Programmer’s Reference Manual
6.11
The increase in the number of pages is not “nearly 50%”, as I said last
month, but is in fact over 70% (excluding the indexes, there were 1,882
pages in the RISC OS 2 manual and there are over 3,200 in the new
edition.) Sorry about the mistake!
6.11
Arm3 hare and tortoise (again)
6.11
(Archive 6.6 p59, 6.8 p30 and 6.10 p63) With the help of Jim Nottingham,
I have found out the source of my problem. The Run file of the Aleph One
!Arm3 application should read:
6.11
IconSprites <Obey$Dir>.!Sprites
6.11
|Use only one of the lines
6.11
|below, comment out the other.
6.11
|Load <Obey$Dir>.CMdlUnix
6.11
Load <Obey$Dir>.CacheMdl
6.11
Load <Obey$Dir>.IconMdl
6.11
<Obey$Dir>.Kick CacheIcon
6.11
In my file, the CacheMdl line had somehow got REM’d out. I vaguely
remember that when I first tried to install !Arm3 with RISC OS 3, I read
the instruction about “using only one of the lines below” as meaning
that only one of the following three lines should be used and, since I
assumed that I no longer needed “CacheMdl” because RISC OS 3 now had a
module to deal with this, I removed it as well as the CMdlUnix line. On
looking at the file again, I now presume that you are only supposed to
remove one of these two lines, so I have reinstated the CacheMdl line
and everything seems fine. A case of trying to be too clever, it seems.
6.11
Redraw problems (and an alternative)
6.11
(Archive 6.8 p29 and 6.10 p63) Richard Hallas confirms that the problem
of incorrect screen redrawing in Edit and DeskEdit is indeed due to a
bug in the Shared C Library and that the problem seems to occur with
particularly long lines. He says that any program which uses the C
library’s text object feature will suffer from the bug until Acorn
corrects it in a future version of CLib. He adds that Acorn has been
aware of the problem for some time, so hopefully a fix should not be too
far off.
6.11
Richard used to use DeskEdit to produce files for input to PMS which
frequently involved very long input lines indeed (up to several hundred
characters). He would occasionally reposition the caret and start
typing, only to find that his input had gone into a separate part of the
file which was not on the screen! Partly because of this bug, he has now
moved over to using the public domain text editor, Zap, which he
thoroughly recommends: “It has its own redraw routines, which are
unbelievably fast and totally accurate: plus you can get twice as much
text on the screen in multisync modes.”  A
6.11

Hints and Tips
6.12
•    Cfs and the cluttered iconbar − There are five drive icons on my
iconbar, together with five Cfs (i.e. Computer Concepts’ Compression)
icons, so there is hardly room for anything else!
6.12
Since version 1.17 of Cfs, however, things have rather changed for the
better, due to the key combination <ctrl-insert> which lets you switch
between the Cfs and the normal window.
6.12
I now find that one Cfs icon on the iconbar is all I need. What I have
done is this: My !Boot file contains nine lines concerning Cfs:
6.12
<d>.!Cfs
6.12
(This runs the application. <d> gives the path for the program, as
described in Archive 3.9, p.8)
6.12
RMEnsure CFSModule 1.17 RMLoad <d> .!Cfs.CFSModule
6.12
RMEnsure CFSModule 1.17 Error CFSModule not loaded
6.12
(These two lines are just to make doubly sure.)
6.12
CFStemp <s>!Scrap.ScrapDir
6.12
NewCFS ADFS::0
6.12
NewCFS ADFS::intern
6.12
NewCFS SCSI::SCSIDrive0
6.12
NewCFS SCSI::SCSIDrive4
6.12
NewCFS SCSI::extern
6.12
(The last six lines duplicate the way in which everything was first set
up from the Cfs iconbar icons.)
6.12
With these commands in the !Boot file, it is now possible to delete all
but one Cfs icon from the iconbar and save the changes. Now one of my
drives (the one I use most often) can have its Cfs directory windows
opened directly. On all the others, I open the relevant uncompressed
window, then press <ctrl-insert>.  Jochen Konietzko, Köln, Germany
6.12
•    Impression arrows − Left and right arrows in Impression are extremely
easy to generate. You start with “<−−” or “−−>” and use <ctrl-R>
inbetween the characters to reduce the spacing (kerning) between the
characters so that you end up with “<−−” or “−−>”. Those two needed one
<ctrl-R> between the dashes and nine <ctrl-R>s between the angle bracket
(inequality sign) and the dash but it will depend on the font size.
(These “dashes” are produced in Impression with <shift-ctrl-hyphen>.
Actually, they are supposed to be minus signs, I think but I prefer them
to the “proper” dash which is <alt-152>. Certainly when I tried the
“proper” dash for this purpose, it produced “—>” which isn’t quite as
good as the height is different and it doesn’t hit quite at the centre
of the angle of the hairpin.) You can see it better at a bigger point
size:
6.12
Two minus signs: “−−>” One dash: “—>>”
6.12
In Times font, which I am using here, the long dash is slightly lighter
and doesn’t match the weight of the angle bracket. However, this trick
does depend on the particular font used − it’s a ‘suck it and see” job.
Ed.
6.12
•    Multistore alphabetical sorting − I recently created a large database
for my church roll in which there was a single name field presenting the
information in the form “Sue Adams”. I set up an index on this field
using the Surname function only to discover that it persisted in sorting
“Sue Adams” after “Margaret Adamson”. An examination of BaseLib (found
by shift-clicking !Multistore and dragging into !Edit) revealed how the
Base Function FNsurn worked and led to the idea of creating another Base
Function, FNchristian, to select the Christian name and any initial.
Using !Edit, I adapted FNsurn and then saved it as a further Base
Function in BaseLib. Then, using the Sort facility with the expression
“FNsurn(@ Name)+FNchristian(@Name)”, the names were sorted into correct
alphabetical order. It is necessary to quit and re-load Multistore to
ensure that it knows of the new Base Function. FNchristian is as
follows:
6.12
DEFFNchristian(A$):LOCALI%,C%
6.12
WHILERIGHT$(A$,1)=“ ”:A$=LEFT$(A$) :ENDWHILE:I%=LENA$: C%=ASCMID$(A$,I%)
6.12
WHILEI%>0AND((C%>64ANDC%<91)OR(C%> 96ANDC%<123)ORC%=45):I%-=1
6.12
C%=ASCMID$(A$,I%):ENDWHILE
6.12
IF INSTR(A$,“ ”)>0THEN=LEFT$(A$, I%-1)ELSE=“”
6.12
Colin Brockie, Kilmarnock.
6.12
•    Multiple stories on one page in Impression − It can sometimes be
extremely desirable to have two or more independent stories on a single
page in Impression. Normally, when you enter text in a single-frame page
and the text fills the frame, a new page is created and the text
continues to flow onto the new page. Creating a second frame on any
given page simply makes a local frame which does not exhibit this
ability to flow text onto a new page when it is full. However, if it
were possible to have two or more frames on a page which could
independently create new pages when full, this would be very useful − it
would make for easy creation of footnotes, for one thing.
6.12
I was recently asked how to achieve this effect by a user who was
setting opera libretti, and who wanted to have two columns side by side
for different languages. A two-column master page layout did not achieve
the desired effect because the text flowed from one column to the next
on the same page, rather than from any given column to the equivalent
column on the following page.
6.12
There is, in fact, a way of achieving the required effect in Impression,
but the technique is a little subtle.
6.12
How to do it − Either choose a suitable master page which exists already
or use the ‘New master page’ menu option to create an appropriate page
with the desired number of columns on it. Create whatever extra frames
you need on your master page in the normal way.
6.12
Now use the ‘Show flow’ option from Impression’s iconbar menu to see how
the text will flow between the frames. Under normal circumstances, to
link frames into the text flow sequence you click <select> on the first
frame in the sequence, followed by <adjust> on the succeeding frames in
the correct order. However, it is not possible to remove frames from the
sequence using this method and so, if you try to alter the order of flow
which is automatically assigned on the master page, you will find you
are not able to do so. Therefore, the only way to remove a frame from
the text flow sequence is to actually delete it and then recreate it.
6.12
So, for each frame which is to contain its own page-creating story,
first select the frame and then open the ‘Alter frame’ dialogue box with
<ctrl-f10>. Make a note (mental or otherwise) of the X and Y coordinates
of the frame, close the dialogue box and press <ctrl-X> to cut the frame
to the clipboard. Immediately you have done this, press <insert> (or
<ctrl-V>) and paste the frame back to where you cut it from. Finally,
open the ‘Alter frame’ box again and restore the coordinates to what
they were before.
6.12
On the face of it, this seems pretty pointless. All you have done is
delete a frame and then put it back again in the same place. However,
the practical effect of the action is to un-link the frame from the text
flow sequence. If you choose ‘Show flow’ from the iconbar menu again,
you will see that the frame no longer has an arrow pointing to or from
it. (In fact, if you have a layout of three or more columns, you will
see that the flow has bypassed this frame, but still connects the
others.)
6.12
What this boils down to is that you can have as many individual stories
as you want on a single master page. Each story can consist either of a
separate frame or any number of linked frames. Any such set of frames on
an Impression master page will cause the creation of a new page when
filled with text in the document.
6.12
The layout shown in the illustration is not intended to be a serious
example but it merely illustrates the idea. The frames for both Story 1
and Story 2 have been linked, whilst the frame for Story 3 is a separate
entity, as is the frame for footnotes.
6.12
As for practical examples of this idea, one has only to look at the old-
style (RISC OS 2) Acorn manuals, such as the BBC Basic Guide or PRM.
These manuals all have a vertical line about a third of the way across
each page, to the right of which is the main text. The headings and sub-
headings all appear to the left of the line. Although the creation of
such pages by conventional methods would be perfectly possible, having
two independent stories (one for the main text, another for the
headings) would greatly simplify the process.
6.12
A similar idea commonly found in educational material is to have a
column, at the right hand side of the page, which contains comments
relating to the ideas presented in the main body of the text.
6.12
Another good use for this idea is the presentation of tabulated material
such as a printed index. If you want a column of numbers to be in a
different style to the reference text the numbers accompany, it is far
easier to have two independent stories than to apply a specific style
for each number on a line.
6.12
When using this approach, there is just one thing to watch out for. If
there is not a direct one-to-one correspondence between the number of
lines in the stories (which there may or may not be, depending on the
document), you must remember to press <ctrl-G> (which corresponds to the
‘Frame=>Force to next’ menu option) when you have typed in all the text
for the shorter story on each page. If you don’t do this, you may find
that, when you position the caret in the frame in which you want to
start typing, the text appears on the previous page. However, you will
soon get used to this.
6.12
As for footnotes, if you know that none of your footnotes will exceed a
single line in depth, you should create your footnote master page frame
to be just large enough to contain one line of text. Then, each time you
press <return>, the caret will advance one page. If you need longer
footnotes, however, make the master page frame large enough to hold the
largest note and remember to use the <ctrl-G> method.  Richard Hallas,
Huddersfield.
6.12
Is it possible to have two or more independent stories running side-by-
side in Ovation? If so, can anyone explain how? Ed.
6.12
•    Printing to files − I experienced the same problem as Tom Rank
(Archive 6.9 p63) with printing to a file. This involved an address
exception being caused at the end of each print run which prevented the
last few graphics lines from being printed. The problem is caused by
ADFSBuffers. Setting this to zero using:
6.12
    *Configure ADFSBuffers 0
6.12
solves the problem.  Paul Mellor, Derby.
6.12
•    Vector overlays − The new version 1.10 of Vector offers the option to
draw lines with overlays (i.e. between 1% and 99% of the line’s interior
is drawn in a different colour), a feature which makes drawing little
map sketches very simple.
6.12
A problem arises if you want to join two roads or two rivers without a
seam.Vector produces something like this (overlay 50%):
6.12
For a decent map you need, instead, something like this:
6.12
If, as in this case, all paths have the same width, it is simple: Just
select them all and merge them through the “Special” menu. (This also
saves memory: The five paths in the first picture take up 1071 bytes,
the merged path only 883 bytes.)
6.12
If the width differs, the thickest line will force its width upon all
the others, making a merge inadvisable; then you will have to apply
little “beauty patches” in the form of completely white rectangles to
cover the unwanted black lines.
6.12
While you rotate the rectangles to align them with the road, it is best
to switch the amount of detail shown on the screen from stage 5 to stage
3, so that all the rectangles appear just as thin black outlines (even
if they are not selected) and this makes an “invisible” shape rather
more easy to handle.
6.12
When you go about rotating those rectangles, it makes sense to lock
everything else first with <ctrl­H> because else it is very simple to
rotate something else besides the little patch.  Jochen Konietzko, Köln,
Germany.  A
Using RISC OS 3.1
6.12
Hugh Eagle
6.12
SCSI problems
6.12
Barry Thompson found one day that he was unable to access the hard
drive, an Oak Worra Winnie 45Mb on an A340 with RISC OS 3.11. The error
message, ‘Disc not understood – has it been formatted?’ came into view.
On Oak’s advice, he refitted RISC OS 2 and the message became ‘Bad free
space map.’ Adrian Look’s !FixMap was unable to help, so the disc had to
be reformatted and, of course, all the data was lost. Someone at Oak
suggested to Barry that SCSI can be less reliable with RISC OS 3 than
with RISC OS 2 and recommended doing a *Checkmap every couple of days
and, if errors were reported, to back everything up and reinitialise the
disc.
6.12
Synonymous pathnames
6.12
Mick Day writes, “I wrote some time ago that the operating system did
not check for synonymous valid references to the RAMDisc pathnames. This
is, in fact, more general. If full and accurate pathnames are always
used, this is no problem; and where removable media are concerned, it is
probably necessary. However, comparison of the ‘Bytes free’ or a CRC
number could have largely eliminated this. Alternatively, if an
abbreviated Disc Spec was given, the filer could have filled it in
completely and thus discovered identical manifestations of filer
windows! It does no harm, but is confusing. If you set the pathnames for
your harddisc in the following short program:
6.12
*Filer_OpenDir RAM::RamDisc0.$
6.12
*Filer_OpenDir RAM::0.$
6.12
*Filer_OpenDir RAM:$
6.12
*Filer_OpenDir SCSI::HardDisc4.$ | set these
6.12
*Filer_OpenDir SCSI::4.$| names to
6.12
*Filer_OpenDir SCSI:$| your system
6.12
*Filer_OpenDir ADFS::MCDay.$
6.12
*Filer_OpenDir ADFS::0.$
6.12
END
6.12
you will find, on running, three identical filer windows on the screen
for RAMDisc, three for Fixed disc and two ADFS floppy! I think this is
sloppy implementation, akin to not checking the disc before issuing
‘Free’!”
6.12
Top bit set characters
6.12
Mick Day uses a Star LC10 mainly for address labels. The dip switches
set it to the Epson character set. However, he sometimes needs to use
top-bit characters in foreign addresses. He thus has implemented a
complete look-up table for the characters 160-255, trying as near as
possible to get the Latin1 set. He says that, interestingly, it is
possible to dodge about and use the IBM set in addition to the Epson
set! The Basic program $.Latin1 on the monthly program disc summarises
it; anyone with a Star LC10 should be able to get it going. It might
work on some others as well.
6.12
Form feeds
6.12
Mick Day also writes: “I have spent a lot of time trying to stop the
PDriver module or its front end from issuing formfeeds. I would rather
decide for myself when the paper is to be thrown out onto the desk. Am I
missing some glaring instruction or is it really impossible to configure
this simple option? I experimented further with !Printers after reading
Tim Nicholson’s article (Archive 6.10 pp 43-47), but got no further. I
suspect that the formfeed is done by PDriver; I have set the printer
configure to the start sequence instead of the end and removed the
formfeed character, but all to no avail. It means that if you drag a
single address to print on labels, you need a specific Printer with a
page length of 9 − a sledgehammer to crack a nut! I want one driver for
the printer which does what I tell it. Can anyone tell me definitively
if the terminal formfeed is configurable or not? Otherwise I could be
experimenting until I die!”
6.12
Testing a printer’s on-line status
6.12
(See Archive 6.11 p51) Mick Day says that he prefers the following
routine to the one published last month (which he says is slower and
escape can interrupt it):
6.12
DEF FNCheckPrinter
6.12
*FX 5,1
6.12
*VDU 2,1,0
6.12
SYS “OS_Byte”,152,3 to ;f%
6.12
f%=f%AND2
6.12
*FX 21,3
6.12
VDU 3
6.12
*FX 5,5
6.12
=f%
6.12
(I’m no expert in these matters, but I believe that the *FX 5,1 only
applies if your printer is connected to the parallel port and the *FX
5,5 might have unwanted effects if your system is set up differently
from Mick’s – HE.)
6.12
He adds that the important point to emphasise is that “the fundamental
difference between earlier Acorn systems and RISC OS 3.1 is that, in
RISC OS 3.1, once you have issued a VDU 3 or equivalent, you are dead!
You can only switch the printer on or press <esc> (if it is enabled).
Page 18 of the Release Notes specifically refers to freezing when
attempting to print to ‘non-existent’ printers. They could have been
much more helpful here in only a few words. I wonder how many man-months
of time-wasting this has generated?”
6.12
AZERTY keyboard
6.12
Mick Day says that, try as he might, he cannot get an AZERTY keyboard
with *Country/*Language/*Keyboard France. He knows of at least one
Belgian colleague who was impressed by his Archimedes but who did not
buy one because he could not get the keyboard he wanted. (We have a
couple of modules that get round this problem − for French and German,
anyway − they are on this month’s program disc.)
12
The following are taken from extracts from recent Acorn Customer Service
Newsletters which Barry Thompson has kindly forwarded:
6.12
RISC OS 3 CMOS problems
6.12
There have been a number of reports of systems not powering up properly.
Acorn suspect this is due to rogue CMOS RAM settings which occurred
while the system was used with RISC OS 2 and they “strongly recommend
that a ‘Delete-power-on’ is performed before the RISC OS 2 upgrade is
removed. This will ensure that CMOS RAM locations now used by RISC OS 3
are set correctly.”
6.12
First Word Plus ‘Window Definition’ error
6.12
If First Word Plus issues the message ‘Window definition will not fit’,
a ‘Delete power on’ will cure the problem (but may well, of course,
corrupt other CMOS settings)!
6.12
Maximum DOS partition sizes
6.12
DOSFS only supports DOS hard disc partitions up to 32Mb. If you create a
larger DOS partition, it will not be possible to access it from the RISC
OS desktop environment.
6.12
DOS partitions larger than 32Mb can only be accessed from within the PC
card or PC Soft (i.e. PC emulator) environment. Therefore, if you wish
to transfer files between a large DOS partition and a RISC OS based
filing system, this should be done via floppy disc(s) or another
(smaller) DOS partition.
6.12
Windows driver with the Acorn 486 card
6.12
Following on from the last point ...
6.12
The Windows Driver disc supplied with the PC card is in Acorn ADFS
format. So, if you have created a DOS partition which is larger than 32
Mb, it will not be possible to use the procedure outlined in the PC Card
User Guide to copy the file on this disc to the DOS partition. Instead,
you should use the following procedure:
6.12
– From within the RISC OS desktop, format a blank floppy disc to 720Kb
DOS format.
6.12
– Click <select> on the floppy disc drive icon to open a Filer window
showing the contents of the root directory of the DOS format floppy (at
this stage it will be empty).
6.12
– Remove the DOS floppy and insert the Windows Driver disc. Click
<select> again and copy the file ARMVGA/DRV from the Windows Driver disc
to the DOS floppy.
6.12
– Start the !PC Card  application.
6.12
– Insert the DOS floppy again.
6.12
– At the DOS prompt type:
6.12
copy a:*.* C:\WINDOWS\SYSTEM
6.12
to copy the correct Windows driver for use with the PC card into the DOS
partition.
6.12
– Load the \WINDOWS\SYSTEM.INI file into the DR DOS text editor by
typing at the DR DOS prompt:
6.12
C:\DRDOS\EDITOR C:\WINDOWS\ SYSTEM.INI
6.12
– Use the cursor keys to move down to the line that reads:
6.12
display.drv=vga.drv
6.12
and change it to read:
6.12
display.drv=armvga.drv
6.12
– Save the modified file by holding down <ctrl> and pressing <K> and
then <Q>. Then press <escape> to return to the DR DOS prompt.
6.12
– The new driver will be loaded the next time you start Windows, and a
different hourglass shape will indicate its presence.
6.12
MEMC1a or no MEMC1a
6.12
The following short command can be used from within Basic to find out
whether a MEMC1a chip is fitted:
6.12
PRINT ?274 AND 1
6.12
If the result is 1 then a MEMC1a is fitted.
6.12
RISC OS 4 Wish List (continued)
6.12
The following are from Richard Burnell:
6.12
•    Shutdown – Include an “are you sure” window after you press <ctrl-
shift-F12> with an option to cancel the Shutdown, in case you press it
by accident.
6.12
•    Unlimited directories – Allow more than 77 files in a directory.
6.12
•    Filer windows – Allow file icons to be positioned where you want them
(as on an Apple Macintosh).
6.12
•    Other disc formats – Include another extension to the filing system to
allow the reading and writing of Amiga and BBC B discs.
6.12
•    Outline fonts – Allow unfilled outlines to be printed (without having
to convert them to paths in !Draw first) as you can in Windows on PCs.
6.12
•    !Draw and !Edit should warn you when you try to save files with names
that already exist.
6.12
•    !Draw – When you drag selected objects, it would be helpful to see the
shape of the objects moving (as you do in Artworks), not just the
bounding box.
6.12
•    !Edit – should allow the option not to strip line numbers to be saved
in the configuration.
6.12
•    Printer buffer – Provide one that allows background printing. (Richard
observes that this was advertised as a feature in the pre-release
RISC OS 3 reviews and finds it surprising that Acorn haven’t yet
released “a working printer buffer module”.)
6.12
Finally, a point from Roger King. A similar suggestion has been made
before but it is worth repeating:
6.12
•Distinguishing non-ADFS discs – When a non-ADFS disc (e.g. a DOS disc)
is mounted, the disc drive icon on the iconbar should change to a
different colour. (Some form of indication like this is particularly
important because the directory names in Filer windows for DOSFS, as for
other “image filing systems”, are prefixed ADFS!) Roger receives many
discs through the post and increasingly often he finds that the discs
are DOS format, presumably because people buy them pre-formatted. The
great snag in using DOS discs in the Acorn environment is the shorter
length of filename allowed. Roger often checks and modifies files sent
to him, then saves them back to the same disc with a “U” (for “Updated”)
tacked onto the end of the filename. If the original file has a name 8
characters long and is on a DOS disc this results in the original file
being overwritten by the updated file since the “U” is discarded from
the filename!  A
6.12