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

[Brian Cameron <Brian Cameron Sun COM>]

Alan:

> Perhaps the Project Ridley developers have clearer intentions of when they
> might want to declare GTK 3.0 and it would provide a clearer break off
> point for Gnome 3.0 but again I would really like to know what the plan
> is if there actually is one.

Although bumping the major version number purely for marketing reasons
isn't necessarily a bad thing, it is important to remember that a
major version bump is an opportunity to improve and simplify the
exposed interfaces.  If there are opportunities to make the API better,
these should be done in conjunction with marketing goals.  Also,
keeping the major number stable shouldn't be viewed as a bad thing.
It highlights stability, and the fact the minor number keeps bumping
up indicates the project is healthy and moving forward.

As stated on the ProjectRidley website on live.gnome.org, a major goal
of the project is to fully deprecate libgnome, libgnomeui,
libgnomeprint, and libgnomeprintui.  Others are also targeted, but I
think these four are probably the most important.  I believe much of
the GNOME stack is already ported away from using libgnome and
libgnomeui, and the print stuff is being merged into GTK+.  Replacing
libgnomecanvis with something that uses cairo would probably also be a
good idea.  Once this work is done, talking about GNOME 3.0 would make
more sense, in my opinion.

Also, it has been discussed that a major release number is a good time
to clear away deprecated functions from the base libraries, moving them
into backwards compatibility libraries for those applications that
need them.  This is just good housekeeping.

It would be ideal, I think, if GNOME 3.0 were a release to highlight
that GNOME has a full set of interfaces for integrating with the
desktop.  Including Platform library API and all the interfaces exposed
in the new and improved Sysadmin Guide.  I think GNOME has a pretty
good set of interfaces already in place.  Though, the print API moving
into the GNOME Platform will be a big step.

However, we should ask ourselves what additional interfaces are needed
to provide the level of integration to make GNOME successful.  Some
ideas to consider:

+ Allow editing of the root menu (right-click menu)
+ Allow applications to install launchers
+ Better KDE/FreeDesktop interoperability?
+ Do HAL/D-BUS (or any others) belong in the GNOME Platform?  Do they
  need work (e.g. API docs) to be ready to be put in the Platform?

If the goal is to make the release announcements more exciting,
perhaps GNOME could do what Solaris does.  Solaris 10 is really
SunOS 5.10 and Solaris 9 is really SunOS 5.9, etc.  Sun doesn't expect
to change the major number anytime, so we just announce the minor
number to the public.  Instant extra excitement.  :)

Brian