Re: GNUstep roadmap (was Re: [Suggestion] GNUstep-test for quality control)

[Philip Mötteli]

Am Mittwoch, 22.10.03, um 10:35 Uhr (Europe/Zurich) schrieb Dennis 
Leeuw:
> To me GNUstep is a development environment, and with that comes the 
> trouble. If you want a real user friendly, crossplatform environment 
> you need to supply the tools for the developer.

That's exactly my opinion.


> There is also the legacy of NeXT and the mirror we keep on looking in 
> called MacOSX.

Well, as long as Gnustep is an OpenStep implementation and OpenStep was 
defined by NeXT…


> One of the biggest things I have seen emerge from/for GNUstep is 
> Renaisance. I think this is/could be the real "pro" of GNUstep 
> compared to other development platforms.

Rennaissance is great, but doesn't make the difference. Unfortunately.


> The real issue here is trying to let go on the "stick to NeXT" and "Oh 
> MacOSX has" ideas.

I don't agree. I actually think exactly the opposite. MOSX is the main 
source for new members of Gnustep. Apple brought Objective-C and Cocoa 
(a "extended2 OpenStep) to the broad public. Many people heard for the 
first time of Gnustep, through Cocoa. A lot of interest in Gnustep is 
due to being crossplatform MOSX->Linux->Windows. So by not following 
MOSX, we loose one major platform. We would be much less 
cross-platform, except, if we implement and maintain the whole Gnustep 
again on yet another platform. But then, where do we get the manpower?


> To summon things up:
> Create a tightly integrated development environment aimed at being 
> crossplatform. This means you have to create a "self contained" 
> environment.
> - GNUstep should have it's own Window Manager (preferably integrated 
> into the backend).
> - GNUstep should have a file browser
> - GNUstep should have a URL browser (http/ftp/...)
> - GNUstep should have a helpsystem (man/info/class browser)
> - GNUstep should have a code editor
> - GNUstep should have a project manager (and documentation tool)
> - GNUstep should have a GUI builder
> - Create a stable backend, which contains everything that is needed to 
> make it work.

I don't see why this must be necessary built by our own, in order to 
build cross-platform software? If others have built such tools, why not 
use them? Or complete, what is missing?


> - The development apps should be aimed at being crossplatform. Use 
> Renaisance (sorry Gregory)

Of course.


> I feel GNUstep should not aim at integration with other environments 
> right now. Let's stay our own island in the sea. And build the best 
> development environment around. The applications will come.

The applications will not come! Or, if they come, they will come much 
too late and with much lower quality (though built with the greates 
libraries). Try to catch up e. g. to something like Mozilla or 
Konqueror, writing your own using only Gnustep(derived) tools! I'd say 
two man-years at least. Do we have that manpower? And is it worth the 
little difference to the others? (And in the meantime, the others will 
have advanced again).


> What I have seen happening from the moment I discovered GNUstep to 
> where it is now great things have been accomplished,

You've been saying, that we should abandon the ties to NeXT/Cocoa. Just 
be aware, that the most time consuming part of an OO software, is its 
design. What we did so far, was let the design be done by others 
(NeXT/Apple) and just copy-cat. So we actually saved 40-50% of the 
time. Imagine, where we would be, doing also the design part ourselves…


> thanks to everyone actually writting code,

I would like to join that too!!!


> and this will continue, I am sure.

Yes, the more people are using it, the quicker it will advance.


> I think GNUstep passed the momentum of being at risk of dying.

Yeah, but I think especially, because Apple adopted OpenStep.


Re
Phil