Re: The Path of GNUstep (Was: Re: Gnustep + mac + windows? Possible?)

[Peter Cooper]

On Mon, Sep 23, 2002 at 12:48:14PM -0400, Tim Harrison wrote:
[snip]
> 
> I really don't think that the GNUstep project should be spending time
> developing anything for Windows at this point.
> 
> First off, Windows is not an open platform, and after the debacle
> regarding the VMWare licencing, and the subsequent statements made about
> using applications that are not open source/free software to benefit
> free software/open source projects, it feels a little... err... dirty.

One of the strengths that GNUstep can offer developers - and this was
one of the cool things about OpenStep - was a wide deployment capability.

Giving developers access to a larger number of potential users should
not be sniffed at. Anything that encourages more developer engagement
with GNUstep (more users on more platforms is a very good political
argument for corporate developers trying to move to GNUstep and free
software tools in general) can result in a vigorous developer community
doing things that deliver value to endusers, and a better GNUstep too.

Deploying free software to non-free platforms helps liberate users there
(if you're taking the FSF line), and even allows them to migrate away
to free platforms over time. That's not some kind of dirty aim, surely?

> Also, with the focus split between the Unix and Windows platform, the
> resources are also split.  As we all know, there's still much to be done
> in GNUstep, and if everyone focused on one platform, there could be that
>    much more progress.
 
I can't see that discouraging those people giving their own time to
currently more marginal platforms (Windows, Open/NetBSD etc vs Linux, say)
is actually going to deliver a "finished" GNUstep any sooner. Free software
is often developed by volunteers. Compulsion tends to push them to other
projects ;)

> As a traditional situation, sure, it's nice to see GNUstep running on
> the platforms that OpenStep was designed for, but the original OpenStep
> specification was done by commercial entities with the resources to
> implement on multiple platforms.  Right now, GNUstep seems to have
> problems just compiling and running on OpenBSD and NetBSD.  Shouldn't
> there be more focus on those systems, rather than Windows?

GNUstep compiles pretty well on NetBSD these days. There are a few more
tweaks yet (threading support is still poor with implications for portaudio
and some other things, and perhaps a tiny amount of work cleaning up the
runtime environment - anyone using a.out-based releases anymore?)

I'm readying some patches to get OpenBSD support happening right now. It's
progressing well. But I'm a volunteer, so no guarantees :)

I think these kinds of discussions that have been taking place are good
ones: where GNUstep is going in relation to Cocoa, the OpenStep spec, look-
and-feel as well as supported target platforms. It's a sign that the core
functionality is pretty close to deployment quality, a fabulous achievement!

Regards

Peter