Re: Gnome 2 infinity and beyond [was Re: Tango and 2.16]

[Dan Winship <danw novell com>]

Andrew Sobala wrote:
> And it can *all* be done *now.* That's the point. It's big. It's hard. 
> It's a lot of work. But there's nothing to stop any of these ideas being 
> implemented. And when we have the code to make this all work, we can 
> build something new: the next generation desktop, and call it GNOME 3.0. 
> But we can't keep doing "little" point-release things and pass a future 
> point-release off as GNOME 3.0 - it doesn't work like that. We'd be 
> crucified.

Yes, we can do it now, and people *are* doing it now. And each cool new
feature will get added to some release as it becomes available. And when
the last feature on that list gets committed to CVS, and Alan says "OK,
NOW can we call it 3.0?", people will say "Oh, come on! How can we call
this release 3.0? It's clearly just an incremental improvement over 2.38!"

The only way we can get a major-version-number-bump-worthy release is if
we write a whole lot of cool code and then *don't release it*. We
managed to do that from 1.4 -> 2.0, because we totally rewrote the
lowest-level libraries, so no one could release their spiffy new
pangoized apps until GObject was ready, and no one could release their
spiffy new GObject-based libraries until pango was ready, etc, and the
total API churn was sufficiently large that no one wanted to bother
trying to backport much from 1.9 to 1.4. And so every new feature that
anyone added anywhere got held back, until the entire desktop could be
released together as 2.0.

The only way this is going to happen again is if there is similarly
massive infrastructural change, and I don't see that happening. (Or if
we deliberately held back working features in one app until there were
corresponding features in every other app as well, and I see that
happening even less.) I don't think there will ever be another quantum
leap forward in GNOME functionality between two consecutive releases.
And I don't think that's a bad thing. I'd rather get a new GNOME release
with 1 point worth of improvements every 6 months than wait 2 years to
get a release with 5 points worth of improvements.

So I don't think we'll ever reach reach a point where we could jump to
"GNOME 3.0" and not feel silly. But at the same time, stick with "2.x"
forever is just as silly. So how about we pull a Solaris and call this
fall's release "GNOME 16", next spring's "GNOME 18", etc?

-- Dan