Re: communication/information between forward-looking projects [was Re: Some info (Ref: GSOC 2008 advice)]

["Mikkel Kamstrup Erlandsen" <mikkel kamstrup gmail com>]

On 04/03/2008, Brian Cameron <Brian Cameron sun com> wrote:

Dave:


> Brian Cameron wrote:
>>> Quoting myself:

>> Whether a user would find a GNOME 2.24 renamed to GNOME 3.0 release
>> exciting would probably depends on what version the user was
>> previously using.  If they were using 2.6 (or something similarly old),
>> they might feel the wealth of new features warrants the major release
>> bump.  Users coming from GNOME 2.22 might not feel the same.
>
> Sure. I'm merely pointing out that the wealth of new functionality which
> has come into GNOME since 2.0, 6 years ago, is worth a major version bump.
>
> The 6 month release cycle means it's happened in small steps, and thus
> we have never had the occasion to recognise the big leaps forward.


I do agree, and I also agree that the time to bump to 3.0 is probably
approaching.  I just think that the GNOME community should invest a
bit of effort wrapping up loose ends before doing a 3.0 release.

A bit of effort? I read that paragraph as "we could just add polish, roll the 3.0 tarball, send the press release and kick back". This is probably not what you mean though :-)

Why not a huge effort and a big bang? Consider the publicity and energy the KDE 4 release has given. KDE 4 also enticed lots of new developers which is never a bad thing (except for mailing list servers and email clients ;-)).

More on this below...

>> I am not opposed to doing a 3.0 release, mind you.  I just think that
>> a 3.0 release should involve more coordination than just deciding to
>> rename GNOME 2.24 to "3.0".  In my opinion, a 3.0 release should make
>> some effort to take GNOME to the next level.  It should not be done
>> because a "major release hasn't happened in a long time, and the KDE
>> team did one."
>
> I mostly agree.
>
> I just think that we set the bar too high for major version number bumps.


Perhaps we can find some middle ground.  I do not personally think it
is necessary for GNOME 3.0 to be an overhaul of GNOME, nor is it
necessary to break any backwards compatibility.

Agreed 100%

I do not believe it will be a catastrophe if we introduce small breakages though. After all Qt4 did not send KDE back to the stone age..? Or maybe they where already there... ahem. Let that rest for now ;-P

However, it does seem sensible for GNOME 3.0 to be timed with when
Project Ridley reaches a sane point.  It would be ideal if the GNOME
3.0 API could comfortably deprecate the interfaces that we have been
working hard to achieve.  Most of this work has already been done.
With gio/gvfs now integrating into GNOME, another major hurdle is
being crossed.

It might also make sense to make a bit more headway on possibly
replacing gconf with dconf, and also to make more progress on replacing
the a11y infrastructure with D-Bus before a 3.0 release.  This way
we could also deprecate bonobo and ORBit2, which has been a longstanding
desire.  I think if we were able to wait until this is done, that we
would really have a new developer platform that would shine enough to
make a 3.0 release especially noteworthy.

Integration of DBus and deprecation of Bonobo is one thing. Another thing I really miss in Gnome is "more framework". Let me explain what I mean by "framework" with the universal communication form of a bullet point list. The individual points are not meant for discussion, just meant to underline what  bugs me from time to time:

platform = "only the Gnome platform - desktop is unstable api, so it does not count"

 1) Why is it not dead easy to put a small email field into your app using $platform?
 2) Why do application developers implement way too many system tray icons?
 3) Why do users request apps to have system tray icons?
 4) Why is it not easy peasy to embed multimedia in your app using $platform?
 5) Why can't I do animations without jumping through hoops using $platform?
 6) Why can't Iembed IM funtionality in my app using $platform?

 7) Why can't I manage my contacts and calendar using $platform?
 8) Why can't I easily do advanced querying of all available file metadata using $platform?
 9) Why is it not dead easy to write an applet?
 10) ... Embedded html widget or integrating web services .. $platform?
 11) shall I go on?

Bonus Point) Proprietary software company thinking - "why can't I integrate $our_app better into Gnome?"

While I don't expect 3.0 to solve all these troubles I expect it to be a firm ground on which we can work on these issues in a clean way.

Cheers,
Mikkel