Re: [Gnome-print] Re: GnomeFont state of affairs

On Sat, 17 Jun 2000, Wolfgang Sourdeau wrote:

> La plume légère, à Fri, Jun 16, 2000 at 04:11:00PM -0400, heure d'inpiration,
> Sean Thomas Middleditch écrivait en ces mots:
> > > Miguel de Icaza <> writes: >
> > > > I think we need parity of goals here.  My goal is to make GNOME
> > > > succeed, to make it ubiquitous, and to make it the best desktop
> > > > environment in the world.  So I am doing everything I can to make that
> > > > happen.
> > > >
> > >
> > > That's a stupid goal. The goal should be to make free software
> > > succeed. GNOME is a means to an end. Don't lose sight of that.
> Both goals could be achieved together. When GNOME will be totally integrated
> in the GNU system (which would be a way of separate GNU from UNIX, for the
> better), the GNOME libraries could become as important as the C library.

Still, at the current stage of development Gnome has chosen to support
proprietary systems too. Either such support has to be dropped, or
separate glue libraries have to be written for those, or we need political
decision that now it is the time of other vendors to emulate GNU, not of
GNU to try being compatible with others.

> Therefore, my question is, why don't you just reduce the amount of libraries
> and merge a couple together.
> Every library which was released to due the GNOME project should not especially
> be part of a big libgnome; every library could probably be merged in
> common libraries depending on its "nature".
> Obviously, this "nature" has to be defined.

Modular system is just easier to upgrade and modify. GNU (is Not Unix) at
all, so it should be trivial to port Gnome to some future architecture,
having very different C library.

> But for example, libgnome and bonobo could be merged together since they are
> non-graphical (GNOME) system libraries. GTK should be kept independent since
> it could be used as just an X toolkit, but libgnomeui could include other
> things.
> Why don't we merge the ORBit libraries together... Is it required to have
> libgnorba, liboaf, etc independent of it? Wasn't ORBit designed just for

Still I think, we have to make differences between technical and
ideological dependencies. There is IMHO no inherent evil in having huge
dependency lists, as long as those are handled by few
meta-packages. Strictly speaking, most your applications depend on Gnome
1.2, although techically they depend on 20-50 different libraries.

> Following what I could perceived from some discussions I had with Richard
> Stallman, GNOME is definitely the way to go. (The GNU Projects need to be
> integrated as much as possible...)		

This needs the aforementioned political decision.
Also the GNU projects includes people with very different backgrounds and
experience lists. For many hacking some mega-library is simply too hard to
start with - gettings hands dirty with some little thing is much easier
and the possibility of breaking other things is much smaller.
Just imagine idea changing some function behaviour slightly in libc - I
would dare not to do/experiment with it just to see, what happens. I would
be forced to study most of C library, to completely understand its
functioning etc. If hacking libgnome, I can at least be sure, my console
apps will continue working. Changing libc requires carefully set up build
environment, at least. etc.

My 0.02

[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]