Path: utzoo!news-server.csri.toronto.edu!rutgers!bellcore!uunet!ora!bloom-beacon!dont-send-mail-to-path-lines
From: mouse@lightning.mcrcim.mcgill.EDU (der Mouse)
Newsgroups: comp.windows.x
Subject: Re: choice of toolkits
Message-ID: <9103140712.AA07209@lightning.McRCIM.McGill.EDU>
Date: 14 Mar 91 07:12:06 GMT
Sender: daemon@athena.mit.edu (Mr Background)
Organization: The Internet
Lines: 78

RMS> Some of you may hope to use the GNU system when it is completed.  The
RMS> GNU system will come with all the free toolkits, but it won't come
RMS> with any proprietary ones.

Kee> There are over a dozen companies providing Motif for Suns.  What
Kee> makes you think they won't provide it for GNU either?  Or will
Kee> linking it with the GNU libraries put GNU restrictions on it?

CF> The point is that the GNU project intends to produce a free
CF> operating system.
(I could get snarky here about how "free" GPV-infected code is, but
let's let that pass for now.)
CF> Vendors can port the Motif toolkit or OLIT to GNU, in which case
CF> you can develop Motif or OLIT applications (depending on your
CF> preference) under GNU.  However these applications will not run on
CF> any GNU installation unless the Motif toolkit or OPEN LOOK toolkit
CF> is made available on terms similar to the GNU general public
CF> license.

Say what?  They'll run on any system with the Motif / OL toolkit
installed.  What does this have to do with the terms under which said
toolkit is distributed?

> In summary you can write software to run under GNU O/S using a
> proprietary, for-money, toolkit but you can not write *GNU* software
> unless you use a free toolkit.

Oh.  This applies only those trying to write code *for Project GNU*?
Then of course you are correct, but your statement applies to only a
very tiny fraction of the people it appeared to be addressed to.

> One possibility would be for OSF to donate the Motif toolkit and mwm
> to the GNU project.

I think this would be a bad move on their part.  Once they do this, it
will be more or less permanently infected with the GPV.  The OSF will
then be unable to even ship it with their systems, never mind sell it.
Donating it to the X Consortium strikes me as a much better move, if
they really want to make it generally available.

> [One OL implementation is already free], although its free nature is
> being undermined by the ludicrous prices charged by third party
> vendors for ports (yes, Unipress, I mean you!).

How does this "undermine" its "free nature"?  You want it, you can ftp
it; you don't want to put in the work to port it, you can pay Unipress
for an already-ported version.  Or you can try to find someone on the
net who's done it and is willing to share.  Or you can bite the bullet,
do it yourself, and, if you really believe in its "free nature", make
the result available to the *next* poor sod who's faced with the same
problem.

> [ [IMO] Athena widget set [is not] suitable for writing applications
>   that will be used by ["user" types] [...] [lack of] any real
>   conventions that encourage consistent behaviour (ie. a style guide)
>   and having little aesthetic appeal. ]

I'm not going to touch the bit about aesthetic appeal; one user's
aesthetic appeal is another's intolerable ugliness.

As for style guides and consistent behavior, I feel intelligent
behavior is often more important than consistent behavior.  I'm sure it
would be no trouble at all to build an application that meticulously
conforms to the Motif, or Open Look, style guide, but is nonetheless a
horror to try to use because the design is bad.  Consistent, but bad.
Much like the truism that "there is not, has never been, and never will
be a useful programming language in which it is the least bit difficult
to write bad programs".

"For a foolish consistency is the hobgoblin of little minds" (or
something very much like that)  - Pope, I think it was.  This hue and
cry about "we must have consistency over all else" seems like an
excellent example.  IMO, of course, as are all opinions in the above.

					der Mouse

			old: mcgill-vision!mouse
			new: mouse@larry.mcrcim.mcgill.edu