Re: Some thoughts on the structure of Gnome



fre, 04.06.2004 kl. 20.44 +0200, skrev Steven Van Impe:
> > Um, GNOME is already broken up this way... There is the Platform, which
> > are just the core development bits.  Then there is the Desktop, which
> > are the applications built on the Platform that are essential for pretty
> > much all users to use a computer; i.e., things like the file manager,
> > panels, browser, etc.
> 
> My definition of the core would be an actual core.
> Right now I see stuff like eye of gnome, gpdf, small utilities, applets
> etc in the 'core' (Desktop modules list).

So basic stuff like viewing a pdf or a picture isn't core functionality
in a modern desktop?

> If you accept a broad definition like this than you HAVE to include
> stuff like Totem, Rhythmbox, GST, Evolution and even the GIMP right
> away, and you will end up approaching infinity.

Well, Totem and Rhythmbox have been proposed before together with both
GST and Evolution. So it's no great secret that we accept this broad
definition.

> Why not get ALL of these (expect stuff like nautilus, gnome panel, which
> aren't up for discussion) out of the desktop, and into some group for
> "Gnome based applications".
> This last group of software would not be considered the core desktop, so
> it would allow for more software to be included, with far less
> discussions.
> 
What would this buy us really? It's not like having discussions about
what software to include in the platform is wasted time. Proposed
applications/tools that have been dropped or withdrawn in earlier cycles
have probably had benefits from those very discussions and the same goes
for the general discussion about what exactly constitutes the Core
Desktop (TM)

> > That is where most of the discussions about what to include happen. 
> > i.e., do modern users expect a music player to be a part of their
> > desktop, or is it an optional component only used by a fraction of
> > desktop users?
> 
> I'd say these are common applications FOR your desktop, not your actual
> desktop.
> 
So our views differ...

> > How things are packaged have absolutely nothing to do with understanding
> > GNOME.  Packaging is just how parts are delivered to end users.  Most
> > end users don't use GNOME's direct source packages but instead use
> > distros' specialized packages which may or may not even reflect the
> > structure of GNOME's upstream packaging.
> 
> I don't think you should consider distributions in their current form to
> be of any significance for the future of desktop linux. To me, none of
> them have the right mentality to be accepted by average (Windows) users.
> I have some thoughts for a "real" desktop distribution as well, but
> that's a whole other discussion :).
> 
Well, it's what we're trying to achieve, but it's definitely a different
thread at least :-)

> > That would get ugly.  You'd need to legally enforce usage of the
> > certificate in order for it to have any real meaning.  Otherwise people
> > will start using it for all their apps just to make it more appealing to
> > users even if it completely fails to integrate with the GNOME desktop at
> > all.
> > 
> > Any application that is incldued in the core GNOME desktop would already
> > have this seal (the requirements for being in the Desktop are pretty
> > close to what you'd want to put on this seal), and applications outside
> > the desktop are very problematic.  Each and every release of the
> > application would need to be recertified since even a minor point
> > release might include HIG-incompliant UI, broken translations, bugs
> > hindering proper use of Platform technologies, and so on.
> 
> I have to agree with that. But why not use a similar structure: include
> programs into the "Gnome based applications" (~ certified programs)
> group in the same way as you include software into gnome right now ?
> Just raise the standard a bit so you get a nicely integrated suite of
> software that makes full use of what you provide in the core desktop.
> Again: if you only use a 'core desktop' group, you also can't include
> for example multiple media players or advanced software like the GIMP,
> which you could with a structure like the one explained (all Gnome users
> use the same "core" platform, no questions asked, and can choose their
> favorite applications while being sure they are fully certified,
> although they aren't part of the core desktop).
> 
Isn't this what Fifth Toe is all about? There has been discussion about
a revamp of the "classes" in the past though, and the archives probably
have the details.

> > No, the discussions will turn into "Is this application Approved or
> > not."  And then there will still be a "does this actually belong in
> > Desktop or should it remain a separate third-party application?"
> 
> I think NONE of the applications FOR your desktop should go into the
> desktop, but into a different group with different criteria for
> inclusion
> 
If it's core functionality for every day work it should be in the
desktop, if it's marginal it shouldn't IMO.

> > Everyone always has this option.  Epiphany is included in the Desktop
> > release yet mainstream distros either don't package Epiphany at all or
> > include it as optional software that is not installed by default.  The
> > Desktop release has pretty much zero impact on real users (which is
> > another reason for the arguing about what's included in Desktop and what
> > isn't) because users don't get their software from GNOME, they get it
> > from a pre-packaged binary distributor, whom are free to modify the
> > GNOME package set and sources however they so wish.  If you don't like
> > Nautilus, install Velocity or something else.  Don't like Metacity? 
> > Replace it.  Rather use Firefox instead of Epiphany?  Go ahead. 
> > Evolution/Totem/Rhythmbox not your style?  Nobody's shoving them down
> > your throat.
> 
> Again, I don't think you should involve distributions, but make sure
> Gnome as you release it is a finished product, ready to be used.

You think our average target user is prepared to download tarballs and
install from source? We certainly don't have the resources to build
binary packages for all platforms/architectures ourselves, so I don't
think this is a worthwhile goal for the GNOME project. This is exactly
what we have distros for. If you want to revive the GNOME Packaging
Project you're very welcome to do that of course.

> My biggest problems with distros is that they provide way too much
> software to be used by average desktop users. These users often can't
> make choices, and expect to just click 'install' once and have a
> complete working desktop for which they can install the programs they
> already know.
> Also, distributions provide a bunch of software they call 'gnome
> software' but doesn't work well with the actual desktop.
> 
Complain to them then? :)

Cheers
Kjartan





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