Re: Cocoa/Windows parallel dvlpmt

[Nicolas Roard]

On 2004-02-04 01:33:07 +0000 John Stiles <JStiles@blizzard.com> wrote:

> On Feb 3, 2004, at 4:33 PM, Philip Mötteli wrote:
>> 
>>> Few Mac developers are going to be interested in working on a crappier 
>>> version of something we already have.
>> 
>> Apart from AppKit: Please give me some examples! Or are you just talking 
>> out of the blue?
> 
> I think this is where the Cocoa fans are diverging from the GNUstep fans.
> 
> For me, AppKit is what makes Cocoa magic. I could care less about the Cocoa 
> containers--I am comfortable with STL and don't really get a kick out of the 
> cool features of, say, NSArray. For me, Cocoa without AppKit is entirely 
> uninteresting. I like Interface Builder, and "springy" views, and dialog 
> controls that "just work" like you expect, and control-dragging around to 
> make my own actions and outlets. AFAICS GNUstep is not going to give me a 
> comparable "ease-of-development" as Cocoa/AppKit, or even half as easy.

Well, I think you really didn't try recently GNUstep...
What I prefer (and imho is the most interersting) in OpenStep is effectively the
AppKit and IB. Guess what ? GNUstep GUI and Gorm works !

Sure, GNUstep GUI it's not yet as polished as Cocoa, there are still some bugs,
and some Cocoa classes miss (not relevant from an OpenStep point of view,
but I guess that's relevant for a Cocoa dev ^_^) -- not that's really a big deal
anyway (replacing a drawer by a window ... wow ... that's so hard ...)

But all in all, it works, and it had improved dramatically since 1-2 years 
(thanks
particularly to Alex Malmberg, who did just an incredible work with the Text
System and the graphic backend).

In fact, I'm using a Linux Desktop right now with almost uniquely GNUstep apps..
Sure, as a developer I'm not a typical user, but anyway I have Workspace.app,
Terminal.app, GNUMail.app, TalkSoup.app for IRC, Cynthiune for MP3, Burn.app
for CD burn, MPlayer.app for video, etc.

Telling that "GNUstep is not going to give me a comparable 
"ease-of-development" as Cocoa/AppKit, or even half as easy." is just a lie,
or more likely, a not very informed point of view. In my opinion, GNUstep 
already
gives you a very good development framework -- even if some works need to
be done to improve it, it's not that far from Cocoa.
Gorm just works beautifuly now -- the only devel app that doesn't really works 
is 
perhaps ProjectCenter (at least when I try it from time to time, even if the 
latest 
releases shows some progress). And I agree it would be good to have a working, 
stable ProjectCenter. The fact that it didn't progressed much is mainly because 
GNUstep-make is very well done and many GNUstep programmers don't really 
*need* (or even want) ProjectCenter (many more just prefer to use emacs/vim)...
But if you stopped to ProjectCenter, well, it's a pity, because you missed many 
things.

> If you are a developer who really wants a cool set of containers and stuff 
> like that, then by all means investigate GNUstep. But frankly I think STL 
> will work just as well, and look better on your resume ;) You are welcome to 
> disagree here and I'm sure there are legitimate reasons to prefer either one, 
> but things like containers just aren't cool enough to make me want to use 
> GNUstep.

One thing that you didn't mentioned is the Distributed Objects capacities of 
base :) 
 
> PS I could care less how CoreFoundation works under the covers. The fact that 
> GNUstep is in Objective-C from top to bottom is not a selling point for me. 
> If anything, it makes me think that Apple's implementation will probably be a 
> lot faster :) Objective-C is meant to be easy to write, but I don't care how 
> the guts of Cocoa are written; I just want them to be fast and stable!

That's understandable ..
But as GNUstep Base works anyway, and it's portable. GNUstep-gui also works well
on X11, and there is active development on Windows.

One thing that amazed me is that people don't seem to care about the fact that 
your
beloved Cocoa (or OSX for that matters) could very well be thrown by Apple (or
simply, Apple could just go out of business). It's really incredible to be so 
blind with
the #1 interest of GNUstep -- its free software nature -- while so many 
astounding
things were removed from the market (OPENSTEP, OpenStep/Windows/Solaris, EOF..)

Sure, GNUstep needs some works... but isn't it a bit more interesting 
investment to do
as you are *sure* that it will always be available for you ? I don't know you, 
but 
personnally I prefer to invest my time in something that I'm sure will last.

So in the end, arguing about GNUstep unfinished status is imho a pity -- if you 
want a
portable solution on Windows, it would be better to help us. But the line of 
thought 
just seems to be "take whatever is provided but don't participate". Honnestly, 
it's a
shame. It could be understandable if GNUstep's states was as mediocre as you 
seems
to imply, but it's far from being that. Works is needed, sure, but if that was 
so awful,
I wouldn't be able to use GNUstep gui programs now and port Cocoa apps to 
GNUstep
and vice-versa.

By the way, I'm not ignorant about Cocoa, I have MOSX/Debian on my ibook and I 
really
like XCode. But please, don't be so blind about GNUstep..

-- 
Nicolas Roard