Tricks of the Mac Game Programming Gurus/TricksOfTheMacGameProgrammingGurus.iso/Information/CSMP Digest/volume 2/csmp-v2-008.txt

Text File | 1994-12-08 | 37.9 KB | 956 lines | [TEXT/R*ch]

C.S.M.P. Digest             Thu, 04 Feb 93       Volume 2 : Issue 8

Today's Topics:

    PPC Sessions, or Any ports in this storm?
    Gestalt Selectors in APPL
    One binary for all Macs (even 128K)
    Advice needed



The Comp.Sys.Mac.Programmer Digest is moderated by Michael A. Kelly.

The digest is a collection of article threads from the internet newsgroup
comp.sys.mac.programmer.  It is designed for people who read c.s.m.p. semi-
regularly and want an archive of the discussions.  If you don't know what a
newsgroup is, you probably don't have access to it.  Ask your systems
administrator(s) for details.  If you don't have access to news, there is
no way that I know of for you to post articles to the group.

Each issue of the digest contains one or more sets of articles (called
threads), with each set corresponding to a 'discussion' of a particular
subject.  The articles are not edited; all articles included in this digest
are in their original posted form (as received by our news server at
cs.uoregon.edu).  Article threads are not added to the digest until the last
article added to the thread is at least one month old (this is to ensure that
the thread is dead before adding it to the digest).  Article threads that
consist of only one message are generally not included in the digest.

The entire digest is available for anonymous ftp from ftp.cs.uoregon.edu
[128.223.8.8] in the directory /pub/mac/csmp-digest.  Be sure to read the
file /pub/mac/csmp-digest/README before downloading any files.  The most
recent issues are available from sumex-aim.stanford.edu [36.44.0.6] in the
directory /info-mac/digest/csmp.  If you don't have ftp capability, the sumex
archive has a mail server; send a message with the text '$MACarch help' (no
quotes) to [e-mail] for more information.

The digest is also available via email.  Just send a note saying that you
want to be on the digest mailing list to [e-mail], and you will
automatically receive each new issue as it is created.  Sorry, back issues
are not available through the mailing list.

Send administrative mail to [e-mail].


-------------------------------------------------------

From: [e-mail] (Christopher Kempke)
Subject: PPC Sessions, or Any ports in this storm?
Date: 14 Dec 92 12:24:42 GMT
Organization: Oregon State University, Computer Science Dept.


First a simple question:  Can I assume that PPC Session reference numbers
are unique to a running application?   This isn't explicitly stated, but
I'm almost sure it's true.   Will my application crash and burn in system
11.3 if I rely on this now?

Here's the scoop on why I want to know.   I'm writing some code that uses
(potentially) several PPCPorts.  However, these ports share a common
completion routine for sends, as well as receives, informs, etc.

Now, in the completion routine I need to know what port the send or
receive request originated from in order to know what to do with some
statistics data about the transfer.  No problem, both PPCSEND and PPCRECEIVE
include a userdata parameter for passing information to the completion
routines.

    BZZZZT.  Wrong answer, take another number and try again.  They
    actually deliver the userdata to the OTHER process involved in the
    transaction, where it does me no good at all.

Worse, the ReadParam and WriteParam (why the name change from send/receive
to write/read, guys?) don't include the port reference number.  The
only thing I can consistently get is the session reference number, and
of course there's two big problems with that:  a) it's not documented to
be unique, so I can't necessarily map it back to the port number, and
b) there's no obvious way to convert them, short of building a table myself
as I open the sessions, and c) the session reference number is not even
listed as being a readable value in the completion procedure, so potentially
in a later system it could be corrupted or overwritten by the time it reaches
the completion routine.

Is there really no way to do this?   I _can't_ be the only person who
needs to send data from the application to its own completion routines.
Can I?
I can't easily use global variables, because there are a (potentially)
unlimited number of open ports and pending transactions.

This seems like something that people would need to be able to do fairly
often; what am I missing?  (I'd be happy to look stupid just to get this
code finished :-))

    --Chris

+++++++++++++++++++++++++++

From: [e-mail] (Lawrence D'Oliveiro, Waikato University)
Date: 16 Dec 92 04:21:27 GMT
Organization: University of Waikato, Hamilton, New Zealand

In article <[e-mail]>, [e-mail] (Christopher Kempke) writes:
>
> First a simple question:  Can I assume that PPC Session reference numbers
> are unique to a running application?

I think that session and port numbers are unique systemwide, not just to an
application. I've been doing some messing about with PPC from HyperCard,
and several times I've made some mistake in a script, tried quitting HyperCard
and going back in, and failed to open my port because it was already open.

Remember the PPC Toolbox can be used from interrupt level and INIT code, not
necessarily by an application. I don't think it knows anything about
application contexts; that knowledge belongs to higher layers, like the
High-Level Event and AppleEvent Managers.

> Worse, the ReadParam and WriteParam (why the name change from send/receive
> to write/read, guys?) don't include the port reference number.  The
> only thing I can consistently get is the session reference number, and
> of course there's two big problems with that:  a) it's not documented to
> be unique, so I can't necessarily map it back to the port number, and
> b) there's no obvious way to convert them, short of building a table myself
> as I open the sessions, and c) the session reference number is not even
> listed as being a readable value in the completion procedure, so potentially
> in a later system it could be corrupted or overwritten by the time it reaches
> the completion routine.

I haven't heard of any cases where the system altered a field where there
was only a right-arrow next to its description. I trust this is still true
with the PPC Toolbox. (He said with a straight face.)

> Is there really no way to do this?   I _can't_ be the only person who
> needs to send data from the application to its own completion routines.
> Can I?

One common technique is to make the PPC parameter block part of a larger
structure that you know about. You can put whatever information you like in
the fields following it.

There's another technique I like to use (because it also works in more
difficult situations), but it involves generating code at run-time, which
upsets some squeamish types. :-)

Lawrence D'Oliveiro                       fone: [phone]
Computer Services Dept                     fax: [phone]
University of Waikato            electric mail: [e-mail]
Hamilton, New Zealand    37^ 47' 26" S, 175^ 19' 7" E, GMT+13:00

---------------------------

From: [e-mail] (Joe Zobkiw)
Subject: Gestalt Selectors in APPL
Organization: The World Public Access UNIX, Brookline, MA
Date: Wed, 23 Dec 1992 20:40:12 GMT

I need an application to install a gestalt selector.
I know I need to load it into the System Heap but I
have a few questions and comments.

1) Can I perform a "6-byte-xdef" type trick on this
code? I'm not sure I can since even if I mark the
6-byte resource as sysHeap the code it points to is
still in my application heap.

2) If I place my Gestalt selector function in a separate
CODE resource, can I mark that resource as sysHeap and
have it work properly? This would be WITHOUT any 6-byte
tricks.

3) How can I remove the Gestalt selector once the app
quits? Will

    err = ReplaceGestalt(selector, nil, &oldGestaltFunction);

work or will this cause the System to jump to location 0
the next time someone calls my selector?

What's the easiest way? Thanks in advance! :)

Joe Zobkiw

- --
- -------------------------------------------------------------
joe zobkiw         [e-mail]        aol: aflzobkiw

macintosh.midi.synthesis.c.oop.asm.communications.graphics...

+++++++++++++++++++++++++++

From: [e-mail] (Grobbins)
Date: 24 Dec 92 06:19:14 GMT
Organization: Experimental System Software Investigations

In article <[e-mail]> [e-mail] (Joe Zobkiw) writes:
>1) Can I perform a "6-byte-xdef" type trick on this
>code? I'm not sure I can since even if I mark the
>6-byte resource as sysHeap the code it points to is
>still in my application heap.

It should work so long as the application is running, but since the
application may quit and there is no way to remove a gestalt selector,
you will need at least one totally stand-alone block of code.  So the
6-byte trick won't solve this problem.

(This limitation doesn't apply to things like time manager tasks; since
they can be removed, it is appropriate to simply make the task code a
function of the application, or to put an initial jump instruction in
the system heap, and remove the task before the app quits.)

You can copy a function from the application into a block in the
system heap.  If the function were called MySelectorFunction and
it were followed by a procedure called NullProc, then this would
work:

 mySelectorFunctionSize := Ord4(@NullProc) - Ord4(@MySelectorFunction);
 { mySelectorFunctionSize should be something reasonable, definitely >0 }

 { make space in the system heap }
 mySelectorFunctionPtr := NewPtrSys(mySelectorFunctionSize);
 IF mySelectorFunctionPtr <> NIL THEN
    BEGIN
      { copy the function into the system heap }
      BlockMove(@MySelectorFunction, mySelectorFunctionPtr,
                mySelectorFunctionSize);
      retCode := NewGestalt(kMySelector, mySelectorFunctionPtr);
    END;

This is more fragile than just compiling a stand-alone code resource
(on which you'd call set the SysHeap bit, call GetResource, DetachResource,
HLock; see IM VI 3-43) since the BlockMove method makes the assumption
that the functions in the compiled code are in the same relative order
in memory as they are in the source file.  Check the compiler output
with a disassembler (like ResEdit's code viewer) to make certain that
functions are sequential as expected in the final application code.


>2) If I place my Gestalt selector function in a separate
>CODE resource, can I mark that resource as sysHeap and
>have it work properly? This would be WITHOUT any 6-byte
>tricks.

Yep, that's the canonical way.  Be sure it's locked in memory before
installation, too.

>3) How can I remove the Gestalt selector once the app
>quits? Will
>    err = ReplaceGestalt(selector, nil, &oldGestaltFunction);
>work or will this cause the System to jump to location 0
>the next time someone calls my selector?
>What's the easiest way? Thanks in advance! :)

Rather than try to remove the selector, just make a simple replacement
selector function that returns zero or something else distinct and
strand it in the System heap.


Grobbins              [e-mail]

Usual disclaimers apply.

+++++++++++++++++++++++++++

From: [e-mail] (Dave Radcliffe)
Date: 31 Dec 92 22:30:51 GMT
Organization: Apple Computer

In article <[e-mail]>, [e-mail] (Joe Zobkiw) writes:
>
> I need an application to install a gestalt selector.
> I know I need to load it into the System Heap but I
> have a few questions and comments.
>
> 1) Can I perform a "6-byte-xdef" type trick on this
> code? I'm not sure I can since even if I mark the
> 6-byte resource as sysHeap the code it points to is
> still in my application heap.
>
> 2) If I place my Gestalt selector function in a separate
> CODE resource, can I mark that resource as sysHeap and
> have it work properly? This would be WITHOUT any 6-byte
> tricks.
>
> 3) How can I remove the Gestalt selector once the app
> quits? Will
>
>     err = ReplaceGestalt(selector, nil, &oldGestaltFunction);
>
> work or will this cause the System to jump to location 0
> the next time someone calls my selector?
>
> What's the easiest way? Thanks in advance! :)
>
> Joe Zobkiw

There is glue code on the January Developer CD that, assuming all you want your
Gestalt function to do is return a value, solves all your problems.  It is
called GestaltValue.  The ReadMe file is pasted below:

GestaltValue Read Me
) Copyright 1992 Apple Computer, Inc.
All rights reserved
- ------------------------------------
This document describes the GestaltValue functionality as implemented in
GestaltValue.o.  Interfaces to GestaltValue are provided in
GestaltValue.h and GestaltValue.p.

GestaltValue is a set of routines, implemented as glue, that extend and
enhance existing Gestalt functionality.  For a complete description of
Gestalt, refer to Chapter 3 of Inside Macintosh VI.

Rationale

Existing Gestalt functionality allows developers to put Gestalt to use in
their own applications by registering a Gestalt selector using the
NewGestalt or ReplaceGestalt calls, and providing a function which can
return the appropriate response.

While this is a powerful feature of Gestalt, in practice, it is difficult
to take advantage of for a number of reasons:

1) The supplied function must entirely reside in the system heap.

2) The function should be reentrant and not move memory because Gestalt
(and therefore the function) may be called at interrupt time.

3) The function must effectively remain resident until restart.  This is
a problem for applications which may come and go multiple times.

4) Since the purpose of the function is to return a response value, the
function must know how to locate, or otherwise manage that value.

All of this means a lot of work just to allow code to retrieve the response
value via Gestalt.

GestaltValue addresses all these issues.  It is built to deal with the most
common case, simply returning a 32-bit value.  It safely implements and
installs a single common, shared gestalt function and reports back values
via the existing Gestalt mechanism.  This gestalt function gets installed
the first time the glue gets invoked.  Subsequent calls to the glue
(either from the same application, or a different one) cooperate with
the gestalt function to maintain a table of selectors and values for the
function to report.

GestaltValue will eventually be implemented as a standard system trap, in
which case, the glue will simply call the trap.  Although currently provided
as a standalone object file, GestaltValue will eventually be rolled into
development libraries; watch for it in a development system near you.

All three GestaltValue functions defined below can move memory, so they
cannot be called at interrupt time.  But the shared gestalt function that
GestaltValue installs in the system heap is reentrant and does not move
memory (i.e., it satisfies criteria (2) above), so any values registered
via this mechanism can be safely retrieved at interrupt time via the
standard Gestalt mechanism.

Interface

Pascal:

FUNCTION NewGestaltValue(selector: OSType;newValue: LONGINT): OSErr;
FUNCTION ReplaceGestaltValue(selector: OSType;replacementValue: LONGINT): OSErr;
FUNCTION DeleteGestaltValue(selector: OSType): OSErr;

C:

pascal OSErr NewGestaltValue (OSType selector, long newValue);
pascal OSErr ReplaceGestaltValue (OSType selector, long replacementValue);
pascal OSErr DeleteGestaltValue (OSType selector);

NewGestaltValue and ReplaceGestaltValue are analogous to NewGestalt and
ReplaceGestalt.  They take a standard 4 character OSType value and a new
or replacement value.  ReplaceGestaltValue has no parameter analogous
to the oldGestaltFunction parameter of ReplaceGestalt.  If you desire
saving the previous value, you should call Gestalt for that selector
before calling ReplaceGestaltValue.

The usual restrictions on the selector parameter still apply.  Apple
recommends you use your four-character creator sequence for your selector.
All lowercase letter and nonalphabetic ASCII sequences are reserved for Apple.

DeleteGestaltValue makes a Gestalt value unknown to Gestalt.  Subsequent
calls to Gestalt for that selector will return gestaltUndefSelectorErr.

Return values

GestaltValue requires that the _Gestalt trap be implemented.  You will get
unimpErr (-4) if _Gestalt is unavailable.

NewGestaltValue returns gestaltDupSelectorErr (-5552) if you attempt to add
a new value for an existing selector.  You should use ReplaceGestaltValue
instead.

ReplaceGestaltValue returns gestaltUndefSelectorErr (-5551) if you attempt
to replace a non-existent selector.  You should use NewGestaltValue instead.

Both NewGestaltValue and ReplaceGestalt value may return memory errors if
they cannot allocate sufficient room in the system heap for their code and
data.

DeleteGestaltValue returns gestaltUndefSelectorErr if you attempt to delete
a selector unknown to GestaltValue (such as a pre-defined gestalt selector
such as 'fpu ').
- ----
I hope this helps.  Have fun,

Dave Radcliffe
Apple Computer
MacDTS

---------------------------

From: [e-mail] (A. Zysk)
Subject: One binary for all Macs (even 128K)
Date: 29 Dec 92 20:13:04 GMT
Organization: ZEI Software

This is a two part question concerning the mant versions of
MACS out there.  I've divided the questions into an ENGINEERING
section and a MARKETING section.  Any response to any question
is appreciated.


ENGINEERING SECTION
===================

Is there anyone else out there who is trying to keep
thier applications *absolutely* portable to ALL Macs
including the Original 1984 128K Mac?

I am developing on higher powered Macs and cross testing on the
128K.

As a second phase, I plan to obtain or develop a library of tools
that selectivly enable or disable certain features either available
or not available on a particular Mac.

A good example is the pulldown menu.  On 128K Macs, you can only
append a limited number of menu items to a pulldown menu.  This is because
the auto scrolling feature is not available.

I know that producing such a library will be very painful, but I anticipate
the results will be well worth it.  Do any comprehensive texts exist
on this subject?

MARKETING SECTION
=================

Given each type of Mac,

    How many units of each have been manufactured?

    By country, per type of MAC, how many units are still in use?

        What's the breakdown, in as much detail as possible, of
    what they are being used for.

- - A. Zysk
  ZEI Software
  [e-mail]

+++++++++++++++++++++++++++

Date: 29 Dec 92 23:49:25 GMT
Organization: Royal Institute of Technology, Stockholm, Sweden

In <[e-mail]> [e-mail] (A. Zysk) writes:

>This is a two part question concerning the mant versions of
>MACS out there.  I've divided the questions into an ENGINEERING
>section and a MARKETING section.  Any response to any question
>is appreciated.

If you're a developer, a lot of info is available from Apple
regarding your questinos.


>Is there anyone else out there who is trying to keep
>thier applications *absolutely* portable to ALL Macs
>including the Original 1984 128K Mac?

Not me. Not Apple. Neither MPW nor Think C support the
64K ROM anymore. Also, the 68000 is fading out, but still
hanging in there in the PowerBook 100s.

>I know that producing such a library will be very painful, but I anticipate
>the results will be well worth it.  Do any comprehensive texts exist
>on this subject?

Worth it? What's your market? The 128K I know of is a fish tank;
the 128K sold less than, say, Quadras sell today I think (I don't
have any figures)

There's just no point in supporting pre-plus pre-system-6
machines today, and tomorrow that's gonna be pre-sys-7
non-color machines.


- --
 -- Jon W{tte, [e-mail], Mac Hacker Deluxe --

 -- I don't fear death, it's dying that scares me.

+++++++++++++++++++++++++++

From: [e-mail] (Bill Johnston)
Date: 30 Dec 92 00:35:24 GMT
Organization: University of Delaware

In article <[e-mail]> [e-mail] (Jon Wdtte) writes:
>In <[e-mail]> [e-mail] (A. Zysk) writes:
>
>>This is a two part question concerning the mant versions of
>>MACS out there.  I've divided the questions into an ENGINEERING
>>section and a MARKETING section.  Any response to any question
>>is appreciated.
>
>>Is there anyone else out there who is trying to keep
>>thier applications *absolutely* portable to ALL Macs
>>including the Original 1984 128K Mac?

The engineering side of the question is simple:  yes, it's still
possible to come up with a single binary that runs on Mac 128
through Quadra 950 if one is willing to compromise on features.

Engineering compromises aside, the appropriate question for the
marketing side is not how many 128K Macs were sold, or even how
many are still in use.  I'd want to know how many of the owners
of these machines are active software buyers, and I'd guess
that the answer is "not very many".  If the owners of these
machines aren't interested enough in buying current software
to buy hardware needed to run it, there isn't much point in
spending effort to support them -- especially if it means
compromising the product delivered to more likely customers.
- --
- -- Bill Johnston ([e-mail])
- -- 38 Chambers Street; Newark, DE 19711; [phone]

+++++++++++++++++++++++++++

From: [e-mail] (Bill Coleman)
Date: 30 Dec 92 15:22:56 GMT
Organization: Hayes Microcomputer Products, Norcross, GA

In article <[e-mail]>, [e-mail] (A. Zysk) writes:
>
> ENGINEERING SECTION
> ===================
>
> Is there anyone else out there who is trying to keep
> thier applications *absolutely* portable to ALL Macs
> including the Original 1984 128K Mac?

Well, we kept Smartcom II running on 128K Macs for a while. About 5 years
ago, we abandoned support for 128K Macs. We still kept support for 512K
Macs, though. About 2 years ago, we pulled out support for the 512K Mac.
We may still support the Mac XL (Lisa), but we haven't tested it.

I would suggest that you don't put yourself out to support the 128K Mac.

Apple estimated (3 years ago) that there are less than 50,000 128K / 512K
Macs out there. (Most had been upgraded to a Plus) Of these, by now many
have been put out of service. Of those still in use, few of these users are
buying new applications.

So, if your application happens to run on a 128K or 512K Mac, fine. If not,
I wouldn't sweat it. Unless you have specific clients that need support for
these Macs, It isn't worth it. You're better off spending your efforts
in supporting the features of the newer system software.

- --
Bill Coleman, AA4LR                ! CIS: 76067,2327    AppleLink: D1958
Principal Software Engineer        ! Packet Radio: AA4LR @ W4QO
Hayes Microcomputer Products, Inc. ! UUCP: uunet!hayes!bcoleman
POB 105203 Atlanta, GA 30348 USA   ! Internet: [e-mail]
Disclaimer: "My employer doesn't pay me to have opinions."
Quote: "The same light shines on vineyards that makes deserts." -Steve Hackett.

+++++++++++++++++++++++++++

From: [e-mail] (Sandra Stewart-Cole)
Date: 3 Jan 93 19:30:01 GMT
Organization: University of Illinois at Urbana

The 128 and 512k are not worth supporting. As someone who actually wrote
programs for them I am saddened in saying that, but the simple fact is that the
total number of pre-plus Macs sold was small, and since there were Apple
upgrades available to Plus status, 3rd-party upgrades to Plus status, and some
really hazardously shoddy power supplies, ther are not too many of them left
out there.

I have seen twice in my 8-year history using the Mac a citation of exponential
sales growth. The first was with the Plus, then with the price cut of '90. Both
time the Mac press ended up drooling about Apple having sold as many units in a
year as in all previous years. (I'm being vague because I can't remember the
exact citations, but remember reading such stories)  Given that a used Plus is
a $300 investment, the pre-Plus market is by nature a cheapskate market. That
makes it real tough to actually sell anything to them, and if you try to make
software that can work on ANY Mac you will have to count on selling to people
with modern macs and maybe having some 2nd-hand sales to people with old Macs.
On the other hand, if you restrict yourself to using only that information in
Inside Mac vI-vIII, and add harmless things like vers and SIZE resources for
later Macs to use, it is a SIMPLE project. The only trouble is that you will
have some pretty austere programs.

---------------------------

From: [e-mail] ()
Subject: Advice needed
Organization: computing
Date: Sun, 13 Dec 1992 06:29:36 GMT

A software company is trying to buy an application I wrote for Macintosh.
Since I don't have this kind of experience before, I would appreciate if
experienced people on the net can give me some good advice:

1. Usually what's the percentage the software author can get from the
company's profit? Any estimation in terms of dollar? (assuming say, $70
each copy.)

2. If the company offer fixed percentage of their profit, how am I supposed
to make sure they will keep their promise, since I don't have access to
their balance sheet?

3. Do I need a lawyer to make it legally binding?

4. Any tips? Precautions? Suggestions?

All helps and suggestions will be very appreciated.



+++++++++++++++++++++++++++

From: [e-mail] (John Werner)
Date: 13 Dec 92 08:30:37 GMT
Organization: School of Education, U.C. Berkeley

In article <[e-mail]> [e-mail] writes:

>1. Usually what's the percentage the software author can get from the
>company's profit? Any estimation in terms of dollar? (assuming say, $70
>each copy.)

Don't agree to a percentage of the "profit", whatever that means.  Ask
for a percentge of the gross, i.e. you get a percentage of every
dollar of revenue they get on your program, regardless of their
expenses to produce it.  This way you don't have to worry quite as
much about them fudging their costs to make the profit look lower.  On
one educational program I wrote, I and the other authors get 15% of
the gross.  We also got a decent advance.  In hindsight, we might have
been able to get more.

>2. If the company offer fixed percentage of their profit, how am I supposed
>to make sure they will keep their promise, since I don't have access to
>their balance sheet?

See #1.  If you get a cut of the revenues instead of the profits, you
just need to see the sales reports, not the balance sheet.  The
contract should say that you get to see some sort of sales report, and
should say how often.  I think we get one every 6 months.

>3. Do I need a lawyer to make it legally binding?

You need a good, tightly-worded contract to make it legally binding.
Unfortunately, you this usually means you need a lawyer.  On the
program I wrote, we negotiated our own contract, starting with the
pre-canned one the publisher tried to get us to sign and combined with
lots of advice from our lawyer.  We had him look over a few drafts and
the final contract before we signed it.  I think we ended up paying
him about $500.  This was in Chicago; I can probably find his
name/number if you want them.

>4. Any tips? Precautions? Suggestions?

Make sure you have a valid, registered copyright on the program.
Don't just put a copyright on it, send in the forms to the copyright
office too.

Try to have the contract say that you retain copyright in the program;
it gives you more control and rights, and a better legal position if
you ever end up in court.  If they insist on getting the copyright, be
aware that you're giving something up, and try to get something from
them in return.

Depending on how commercial the program is, you might want to try to
get some sort of promise that they will put a certain amount of effort
into marketing/advertising or the rights revert to you.  This would be
hard to get, but might be something to negotiate away.
- --
John Werner                        [e-mail]
UC Berkeley School of Education    510-642-9651

+++++++++++++++++++++++++++

From: [e-mail] (David Phillip Oster)
Date: 14 Dec 92 08:39:07 GMT
Organization: Whole Earth 'Lectronic Link

In article <[e-mail]> [e-mail] writes:
>A software company is trying to buy an application I wrote for Macintosh.
>Since I don't have this kind of experience before, I would appreciate if
>experienced people on the net can give me some good advice:

>1. Usually what's the percentage the software author can get from the
>company's profit? Any estimation in terms of dollar? (assuming say, $70
>each copy.)
Whatever you can negotiate. IIt has been reported in the trade press that
the author of the Atari version of Frogger made $0.75Million. It has been
reported that Randy Wigginton got $1. for each copy of MacWrite given away
with new macintoshes for the first 1 million macs. (Not bad for 6 months
work.)

>2. If the company offer fixed percentage of their profit, how am I supposed
>to make sure they will keep their promise, since I don't have access to
>their balance sheet?
Many contracts specify that you get periodic access to their balance sheets,
say once a year. What does "profit" mean here? Are their own salaries
included? Do they subcontract production of the boxes to a production
house that is, say, owned by a parent of the publisher, so the cost of
production always equals income, (i.e, the "profit" is siphoned out of
the publisher to the printer, leaving you with a piece of nothing.)?
It has been reported that movie studios do this trick a lot.

>3. Do I need a lawyer to make it legally binding?
No, a contract does not _require_ a lawyer to be legally binding, but a
lawyer is trained to point out spots that another lawyer can use to cause
you trouble. (This process of "sticking an oar in" itself causes trouble
and slows the negotiations down.)  See the next point.

>4. Any tips? Precautions? Suggestions?
No Lo press, Parker St., Berkeley CA, publishes a number of good self help
books on selling software, including selling to a publisher for royalties.
Get them. Read them.

"Getting to yes" and its sequels are good books on the art of negotiation,
and there are many good books on how to sell. In the real world, what you
get is what you get and everything is negotiable, although other parties
often point to a printed sheet of rules as a means of making you think you
can't negotiate.
Remember that you can always hire a trained negotiator, but the better he
is, the better he will be at negotiating his deal with _you_, so there is
a point of diminishing returns.

Remember subsidiary rights, like what happens if the publisher wants a
version on another computer.

Think of it as you hiring a "publisher" to produce, distribute, and
market your work. How much is that worth?

I've been doing this for a living, but I sure don't know it all.
Anyone else have advice?

+++++++++++++++++++++++++++

From: [e-mail] (Ed Gaillard)
Date: 14 Dec 92 00:44:17 GMT
Organization: PANIX Public Access Unix, NYC

In <[e-mail]>
[e-mail] (A Man With No Name) writes:

>A software company is trying to buy an application I wrote for Macintosh.
>Since I don't have this kind of experience before, I would appreciate if
>experienced people on the net can give me some good advice:

[...]

>2. If the company offer fixed percentage of their profit, how am I supposed
>to make sure they will keep their promise, since I don't have access to
>their balance sheet?

Your contract should specify how this is to be audited.  BTW, if you
agree to a fixed royalty per copy (or a percentage of the sales,
rather than the profit), you will still have the problem of
verifying the sales figures and pricing.

>3. Do I need a lawyer to make it legally binding?

Oh, yes.  Definitely.  You _need_ a lawyer.  A little time and money
spect on a lawyer up-front, while negotiating your contract, will save
you a _lot_ of unpleasantness later.

>4. Any tips? Precautions? Suggestions?

One thing to be careful of is, if your program controls special
hardware that this company makes or sells, you must be protected from
their selling the harware and _giving away_ your program.

>All helps and suggestions will be very appreciated.

Disclaimer:  I'm _not_ a lawyer, just someone who has had a bad
experience with this kind of thing.  Get a lawyer, most preferably one
with experience in software royalty agreements.

- --
Ed Gaillard <[e-mail]>
I can't see the lines I used to think I could read between.

+++++++++++++++++++++++++++

From: [e-mail] (Kevin D. Lee)
Date: 14 Dec 92 14:32:32 GMT
Organization: University of Minnesota, Minneapolis, CSci dept.

In article <[e-mail]> [e-mail] writes:
>A software company is trying to buy an application I wrote for Macintosh.

    Don't sell the copyright, just sell the marketing rights in return
    for a percentage of sales.  One way to protect yourself is to require
    a yearly performance figure, if the software sales don't reach the
    minumum sales figure (i.e. generate enough roylaties) then at your
    option, all rights revert to you.


>
>1. Usually what's the percentage the software author can get from the
>company's profit? Any estimation in terms of dollar? (assuming say, $70
>each copy.)
    10-20% seems to be the norm for roylaties.  Make sure that the
    royalty is on net sales.  By the way, roylaties are nice because
    you do not have to pay social security on roylaties.

>
>2. If the company offer fixed percentage of their profit, how am I supposed
>to make sure they will keep their promise, since I don't have access to
>their balance sheet?
    In you contract insist on the right to audit the books. They
    may require that it be done by a CPA.  Also insist that if
    the audit shows they made an error they pay for the audit.

B
>
>3. Do I need a lawyer to make it legally binding?
    No, but you may feel more comfortable if you have a lawyer.
    Try to do as musch of the negotiating yourself and just
    use the lawyer as a refrence.

B
>
>4. Any tips? Precautions? Suggestions?
    There are several books published about negotiating your
    own book contracts.  There are enough similarities that
    it is worth your while to look at them.
    Good Luck!
>
>All helps and suggestions will be very appreciated.
>
>

+++++++++++++++++++++++++++

From: [e-mail] (Jon Wtte)
Date: 14 Dec 92 16:23:29 GMT
Organization: Royal Institute of Technology, Stockholm, Sweden

In <[e-mail]> [e-mail] (Ed Gaillard) writes:

>>2. If the company offer fixed percentage of their profit, how am I supposed
>>to make sure they will keep their promise, since I don't have access to
>>their balance sheet?

>Your contract should specify how this is to be audited.  BTW, if you
>agree to a fixed royalty per copy (or a percentage of the sales,
>rather than the profit), you will still have the problem of
>verifying the sales figures and pricing.

Let the contract specify that you DO have access to their
bookkeeping. You have the program, right?

>>3. Do I need a lawyer to make it legally binding?

>Oh, yes.  Definitely.  You _need_ a lawyer.  A little time and money

And a standard contract shouldn't take more than say 10 hours
for an experienced lawyer. Add the cost of the contract as a
lump sum for the program. (If they don't want to pay, they
shouldn't have the program)

>>4. Any tips? Precautions? Suggestions?

Yes. Don't go for a percentage, go for a fixed rate. Say $3
or whatever per sold copy (plus the cost of a good contract
initially, of course) Better yet, figure out how much time
you have spent on the program, multiply by your hourly rate
(say $50 for a cheap programmer, $100 for a decent one) and
just seel off the source. If they want support, they pay for
it later.

That goes even if you sell it per copy; if they want support
or changes, THEY PAY BY THE HOUR. Be a little tough; either
they want the program and you should be rewarded, or the
program is insignificant, and then you didn't lose much anyway.

Oh, and be sure to make the contract say THEY assume full responsibility
for any demands or suits caused by them selling the program. You
just hand them the code (or duplicated disks, in case of a per-copy
system) and make no warranty.

Of course, I'm no lawyer, just a software designer who fortunately
consulted with a (good) lawyer before signing anything.

Cheers,

                        / h+

- --
 -- Jon W{tte, [e-mail], Mac Hacker Deluxe --

   This article printed on 100% recycled electrons.

+++++++++++++++++++++++++++

From: [e-mail] (Ed Gaillard)
Date: 16 Dec 92 03:54:20 GMT
Organization: PANIX Public Access Unix, NYC

In reply to my post, I received an Email from Sean J. Crist
<[e-mail]>, which follows:

<begin forwarded mail>

>>3. Do I need a lawyer to make it legally binding?
>
>Oh, yes.  Definitely.  You _need_ a lawyer.  A little time and money
>spect on a lawyer up-front, while negotiating your contract, will save
>you a _lot_ of unpleasantness later.

I've tried posting this but for some reason the server keeps rejecting it.
So I'll just mail it to you; I'd like it if you could post it, since I
don't seem to be able.

An important point of clarification:  A contract is legally binding as
soon as it's signed, regardless of whether an attorney was consulted.
What an attorney can do is point out parts of the contract which aren't
legally enforceable, or parts which might have legal ramifications which
you might not have anticipated.  A contract scribbled on a cocktail napkin
is binding, as long as all parties have signed it.

So, do you need an attorney to make a contract legally binding?  No.  Is
it a good idea to consult an attorney anyway?  Yes.

<end fowarded mail>

Mr. Crist is absolutely right.  I'm sorry for any confusion I might have
caused.  I saw "do I need a lawyer" and went off half-cocked.  I meant
"don't sign anything until you get a lawyer!"

- --
Ed Gaillard <[e-mail]>
I can't see the lines I used to think I could read between.

---------------------------

End of C.S.M.P. Digest
**********************