Re: GNUstep Window Manager (was RE: Idea)

[Nicola Pero]

>>>>> "Helge" == Helge Hess <helge.hess@skyrix.com> writes:

    Helge> You say GNUstep-gui isn't an X-toolkit, but
    Helge> doesn't provide a single backend which isn't X based -
    Helge> something is wrong here.  

It took years for gtk to get a windows port after it was working on
gnu/linux.  Give us at least the same time.


    >> Certainly the API is totally un-X-like, which is perhaps a
    >> reason why we have few developers (and a reason why we have
    >> those of us that wqe do have).

    Helge> No, I don't think that this is the reason at all. Eg you
    Helge> also have to learn the gtk+ object system to be able to
    Helge> develop in gtk+. I don't think gtk+ or Qt are in any way
    Helge> 'X'-like.  

Yes, they are - in the sense that Richard means.  We have all our
problems with focusing of windows because the OPENSTEP API goes
against the standard X conventions.


    Helge> For me the reason is that there was no usable GS
    Helge> GUI yet. Before Qt and gtk+ started with a fully working
    Helge> framework, not with a thing in development. 

The first time I used gtk, it had to offer much less working stuff
than gnustep gui offers now.  And, it had nothing comparable to the
gnustep base library below itself <actually, it does not have it even
now>.  Still, people were using it.


    Helge> gs-gui isn't even in this start stage.

I have to strongly disagree - gs-gui is absolutely usable now.  I
wonder why people don't try it.


    Helge> Killer: If GNUstep manages to create a really good text
    Helge> framework, this could be the breakthrough. 

After seeing you say that gs-gui isn't even in a start stage, I don't
think a great text framework would make any change.  If we were able
to complete the effort of implementing a really good text framework
(which - I tell you - is not unlikely if a real effort is done), you
would probably not even notice.

Looks a bit like the table view.  People went on saying that without a
good table view the gui was useless.  So, we took the effort of
implementing it.  But when I had spent my nights coding it, nobody
noticed the table view was there.  Strange, huh ?  People still went
on saying that because we didn't have a table view (while we did then)
the gui was useless.  I even received personal mails about this.  That
was ridicolous and disappointing.

This makes me think that it's not really whether the library works or
not which makes a difference.  The problem is that people are *not*
using the library, and this is not because it works or it doesn't
work, it is for some other strange reason - marketing ?  Well, nobody
forces them to use the library, only it doesn't make sense that they
come here telling us we should trash or rewrite the library because it
doesn't work when they have not even tried it.  <yes, looks a lot like
the recent bashing on window maker for not doing things is perfectly
does>.

I'm not saying the gnustep gui library is perfect or complete - I'm
aware of much more bugs than you could probably notice in hours of
testing - but perhaps that's precisely the point - the library does
its job and can be used.  Yes - editing rich text is the exception,
and has its many evident wrinkles - but - in your comparison - gtk had
it working much much later than its start stage.