Re: GNUstep Window Manager (was RE: Idea)

[Richard Frith-Macdonald]

On Sunday, January 7, 2001, at 02:53 PM, Helge Hess wrote:

> Richard Frith-Macdonald wrote: 
> > I don't think there is any problem with the notion of sharing with other 
> > projects in 
> > principle, it's just not technically realistic due to basic architectural 
> differences... 
> > With many toolkits there is just no reasonable way to share - we would just 
> > have to adopt 
> > the other toolkit and discard our existing work.  The problem with doing 
> > that is that, 
> > our existing toolkits are overall as good as, or better than the toolkits 
> > we might 
> > look at adopting, so while we would gain on some features, we'd lose out 
> > overall. 
>  
> So in what time do you think gstep-gui will reach the same level of 
> stability/feature-completeness as for example Qt ? After doing some 
> *years* of gstep-gui development without a 1.0 result, may it not indeed 
> be a good, even if hurting, decision  to swap to another toolkit ? 
> This approach is taken by almost any scripting language with great 
> success. On the other side Smalltalk's follow the GNUstep approach to 
> make everything in one language with basically no success. 

You seem to have missed the main point ... that we can't easily incorporate the 
best of
other toolkits, but would realistically have to adopt them as a whole!

As for the Qt comparison - that's a bit like flamebait in a gnustep discussion 
list.
Of course, from the point of view of the aims of the GNUstep project, gstep-gui 
had
better feature completeness years ago than Qt does now, and (while time to 
completion
is impossible to give becuse it depends on developers working on the project) 
I'd
judge that gstep-gui could be staable long before Qt could be feature-complete.

> > I looked at the gnome stuff, from a viewpoint of trying to integrate with 
> > gstep-base 
> > and it simply had nothing to offer.  However, special purpose libraries 
> > (like libxml) 
> > can reasonably be integrated, and I'm in favour of doing that where it's 
> > feasable. 
>  
> I think this was already discussed one time in the past ;-) Using glib 
> in gstep-base would at least easen using gtk+ in gstep-gui. 

But at a loss of efficiency in GNUstep apps (or hard work avoiding that) ... it 
would
be much more effective to alter gtk+ to use the base library (if we wanted to 
use gtk+).