The Datafile PD-CD 5/DATAFILE_PDCD5.iso/utilities/t/tornado/Guide/9

Text File | 1995-09-23 | 5.9 KB | 109 lines

Miscellaneous
-=-=-=-=-=-=-

Well, this is simply here, as the name denotes, to stick in all the things
not covered by the other parts of this archive. Or, in other words, to boast
about Tornado!
   Tornado is heavily future-proofed. It has provison for infinite length
filenames, infinite files per directory, and interestingly 32 bit filetypes.
Internally it uses all of these, and 'translates' to and from what is
currently used when required. It can handle multiple processor architectures
as and when they appear.
   It's extremely flexible. Tornado will bend to even the most demanding
applications, and can be customised to a level pretty much unheard of on any
architecture. Routines can be replaced by 'fix-it-up' modules which correct
bugs.
   Tornado currently is written entirely in assembler, and it is envisaged it
will remain that way. Tornado is fast, frugal and sucks as much power from
the processor as it can. It should only use about 200k of module space (I'll
remind you about all the things it does!).
   All of Tornado's code is reusuable eg; a piece of code it might use to put
a routine on the SWI vector will always have a SWI attached to it, so that
other programmers may use Tornado's code. Tornado's code is written above
that of the current norm, with extremely flexible abilities, proper error
handling and also, it's fast.
   Tornado supports multiple users on one computer, and stores config files
in such a way that they are different for each user.
   Also, little desktop nicities will be implemented eg; a hotkey which
cycles through the currently open windows (including RISC-OS ones), bringing
each to the front and giving it the caret. And a hotkey which opens the
directory which the pointer (which is dragging a file into the filer) is
over - thus stopping the really annoying times when you drag a file out to
find the directory you want to save into isn't open. And a hotkey & menu
option to send/take all selected stuff to a certain app each time. This
implements an extremely effective intertask clipboard, and the file will be
converted if necessary.

Finally, some may have heard that it is intended that the RISC-OS desktop be
rewritten to make best use of Tornado's facilities. This is a long way off,
but it may become impractical to have the filing system still running on a
RISC-OS level and everything else running at a Tornado level. Rewrites of the
filer, display selector, and definately task manager will certainly be on the
cards. I have a load of things that will be done if this ever happens, thanks
to the people from c.s.a.*. But for the moment, it's not happening.

Other miscellaneous things
-=-=-=-=-=-=-=-=-=-=-=-=-=

Also some may have heard of the Tornado verification procedure. Essentially,
commercial writers writing code using Tornado will be encouraged to send in
their program for testing, for a fee of course. The program will be
meticulously tested, any problems (if any) noted and the writer informed as
to whether the program has passed or not. From hence, the writer may quote in
all adverts that the task passed the test, that it conforms to certain basic
guidelines. Thus, a program with the test passed with have a considerable
advantage in the marketplace over a non-passed program, as the consumer will
know that a program which has passed the test will be guaranteed to have
certain 'nicities' which other programs may not. Also, copies will be
kept of the report sent back to the writer, and made available for public
inspection so that potential buyers can check out a program before buying.

Far away things
-=-=-=-=-=-=-=-

If Tornado really takes off, and for that it needs to be written first!, it
can be expected that a professional commercial suite of software will be
released which will be based around the VisualXXX available for Windows, but
definately _without_ the crippling disadvantages that those programs have.
Essentially, it will allow tornado apps to be written very quickly, and that
will justify its price tag. Users will still be able to write tornado apps
without it using the shareware editor, but obviously it will be slower and
more niggley.
   Up until this commercial release, that limited shareware visual editor
will also be made available for use by writers not wishing to fork out for
the full development suite. This shareware editor will be maintained and
upgraded on an indefinite basis - however, not to the extent that it would
trample on features found in the commercial version. :-)

Other ideas include building in software encryption/compression in on all i/o
streams. In other words, as all i/o is done by tornado, it will be very easy
to compress/decompress data on the fly.
   Another thing we'll get around to eventually is implementing long
filenames and infinite files per directory. This will be done in conjunction
with the Unix zip suite of programs & TFS, and will work by creating a single
archive in the filecore root directory and then doing all i/o to and from
this archive. Since data can be quickly compressed to 50% almost on the fly
for a floppy disc, it's not out of view that the archive will be a full
zipped compressed archive.

Finally ...
-=-=-=-=-=-

Finally, tornado will remain public domain for all time, and will be made as
accessable to users as possible. Development on tornado will go with what the
users want, not what the writers /think/ they want. Unlike Acorn, we want to
make a difference, a change - rather than constantly refining the same
RISC-OS 2 over and over again, which is ultimately self-defeating. Grumbles
in the Acorn market are growing stronger every day, as acorn users watch
other platforms GUI's get better and better, and yet the RISC-OS GUI stays
pretty much exactly where it is. Increasingly, MCIBTYC arguments are becoming
harder and harder to win. Which, for any acorn-loving fanatic, is enough
reason for violent action! :-)


Cheers,
Niall,
   at [e-mail]
   at Niall Douglas@Fidonet#2:257/501.13
   at Niall Douglas@Riscnet#7:353/1.0