Re: Three Point Zero - Idea Mockups

[Havoc Pennington <hp redhat com>]

On Wed, 2005-05-25 at 21:58 +0100, Mike Hearn wrote:
> 
> I was under the impression that if GNOME 3 ever happens, it's most likely
> to be by re-integrating a fork that went off to (successfully) explore new
> ideas. In other words, there's nothing really to decide - one day if
> somebody produces a version of GNOME that is obviously and fundamentally
> better, that can become GNOME 3. From that perspective any and all blue
> sky thinking is legitimate topaz-fodder.
> 
> I'm just repeating comments I read from various *real* GNOME
> developers (unlike me, I'm just a wannabe ;)

I like Owen's point:
        
        But I'd fight the idea that GNOME 3 is some major milestone or
        rewrite. GNOME 3 is just one step on the way to GNOME 3.2 and
        GNOME 4, and GNOME 4.2 and GNOME 5.

There's no evidence a huge break type of release is needed to justify a
version number of 3.

Look at the highlights of a major version of OS X: 
http://www.apple.com/macosx/

Spotlight - equivalent of Beagle
Dashboard - equivalent of gdesklets
Safari RSS

i.e., we can easily call something GNOME 3 because it integrates a few
cool apps or features. That's how we should think of GNOME 3. (And 4, 5,
6.) Maybe our version scheme should be 3.0, 3.2, 3.4; 4.0, 4.2, 4.4;
etc. or maybe just 3, 4, 5, 6

It feels like there's pressure to break ABI just so we can bump the
version number! If we want to bump the version number let's just do it.

The thing I wrote about a fork is: if you were going to make
_structural_ changes then they should be in a longish-term branch. But
I'm not sure anyone is proposing anything structural. Mac Classic to OS
X may have been a big structural change, but all subsequent OS X
releases have not been.

Well, great example today actually: Maemo is a structurally different
desktop project, though it shares lots of GNOME components. Another
example is Lotus Workplace, which also shares some components such as
GTK.

I'm not sure Longhorn is really all that different from Windows XP, from
a what-it-is point of view. It's exactly the petrified wood analogy;
sure it'll be made of all different code from Windows XP, but it'll
still look like pretty much the same shape of thing from a user point of
view.

BTW, if we make structural changes, they should be around user benefits,
not ABI cleanup.

To me defining GNOME 3 as "yay we get to break everything!" is total
crack in every possible way yet some people seem to have that idea.
"Break stuff! woohoo!" is not a vision ;-)

 - Starting a cool new project with different assumptions and goals to
create user benefits not possible in the GNOME structure, awesome.
(Maemo is the example of the day!)

 - Incrementally adding great user benefits and new developer APIs to
GNOME, awesome.

 - Going nuts breaking stuff just so we can change the version number to
3: huh?

 - ABI cleanup: nobody cares. We can do it, sure. Rename the library,
keep the old one in parallel, put the new one in the GNOME release of
your choice, you're done. It's a non-event.


I guess the way I'm framing this,
 - people should talk about cool new projects we could start
   (like Maemo they could be GNOME-related but not actually GNOME, 
    or they could be branches of GNOME with intent to land 
    as a GNOME release eventually)
 - we should have a great ongoing vision for GNOME itself and 
   continuing cool progress through GNOME 3, 4, 5
 - we should just clean up ABI whenever, as required, and not 
   relate it to the version number

They all seem on-topic for this list, though I think the first one most
asks for the caveat "you should be actually planning/able to start the
project before posting too much text about it" ;-) i.e. we don't need
too many posts about how we should rethink the desktop to be virtual
reality augmented by 20 features the Amiga had in 1934.

Havoc