Re: Future of GNOME

[Tom Tromey <tromey cygnus com>]

>>>>> "Dietmar" == Dietmar Maurer <dm@vlsivie.tuwien.ac.at> writes:

Dietmar> I'm a little bit disappointed about gnome. Although many
Dietmar> people implement programs, it seems that we don't have any
Dietmar> new ideas. Most of the initial proposals are ignored, and I
Dietmar> don't see what makes gnome better than other approaches.

I'm curious to know which proposals have been ignored.

I long ago gave up trying to reply to even a fraction of the mail on
the Gnome list.  But most messages do seem to get at least one reply.

If you mean that proposals are discussed and then not implemented,
well, I agree that that happens.  And I'll even agree that this might
occasionally result in Gnome adopting an inferior solution when a
superior one has already been proposed.  I tend to believe that this
happens because people tend to work on what they find interesting.
Even the best proposal might not spark anybody's interest.

I'm not sure what we could do about this.  I doubt there is anything.

I personally try to implement whatever I think is the most important
thing that (a) interests me, and (b) nobody else is doing.  I might
never get around to some good things simply because our priorities
differ...

Also, I think that many people already have extensive to-do lists.  I
know I do.  This makes it harder for a new proposal to get implemented
quickly, because it must first "beat out" everything else.

Dietmar> I think it's time to rethink the gnome design and state more
Dietmar> exactly what gnome should be. The low-level part (gtk,
Dietmar> libgnome..) is OK, but many other things are unclear
Dietmar> (UI-Guidelines, software component model, ...).

On the contrary, I think many Gnome developers have more-or-less
shared a common view of how Gnome ought to work, and what should be
developed.  For instance, the idea of using CORBA to provide
inter-application glue, and the idea of writing a library (libgnomeui)
to make it easy to write an application have both been around a long
time.

There have been some logistical problems.  For instance, CORBA hasn't
really been used until recently due to problems finding an acceptable
ORB.  This particular problem has in turn caused a delay in the
component model -- probably because nobody wants to design on
something that they can't try out.

I agree that in general these ideas might not be expressed very well.
(I've never looked for descriptions of the plan, since I picked it up
by reading the mailing list over a long period.)

There have also been other problems.  I don't know the precise story
behind the UI Guidelines (and I don't want to know, so please don't
start a flame war about it here!), but I gather that there were some
false starts.  This definitely seems to be an area where good help is
desired.  (Disclaimer: for all I know lots of people are working on
it.  I'm not on the gui list right now.)


Dietmar> I've written two programs where it's possible to use the
Dietmar> gnome-mdi interface. But it seems that there is no common
Dietmar> acceptance on gnome-mdi, although it's in the core libraries.

As others have said, this isn't intentional.  Based on the cvs
commits, I'd say MDI work is going strong.

Dietmar> OK, the panel works, but it provides the same functionality
Dietmar> as one module of my window manager, not much more - features
Dietmar> like a drawer are useless for most people. And the menus are
Dietmar> at least uncommon. I can't see any new idea.

I don't understand what you mean here.  What would you propose we do
about it?

I think there are two important realizations:

1. Hardly anybody agrees about what is interesting.  You might loathe
   the panel, but other people like it and want to hack on it.
   My advice is to ignore everything you don't care about.  This is
   what I do and it seems to work well.

2. Even programs that you might think are utterly useless
   non-innovative fluff can have a positive impact on Gnome.  That's
   because the people developing these programs also help (sometimes
   indirectly) refine the core libraries.

Dietmar> control-center:

Dietmar> The control-center uses a new user interface. This is bad
Dietmar> because it's uncommon.

This seems to contradict your previous comments about innovation.
Is there some other kind of new idea that you'd like to see?

I guess I don't really know where you're coming from.  I don't see any
deep problems with Gnome development (except too few developers and
too little documentation -- and the latter is being worked on).  What
would you change to make things better?

Tom