Ideas for "GNOME 3.0"

[Owen Williams <ywwg usa net>]

Quite often the topic of GNOME 3.0 comes up among the GNOME user
community.  To me, the gist of the response from the developer community
seems like bad PR: "don't hold your breath."  I was thinking about
"GNOME 3.0" recently from a marketing perspective, and what it would
mean if we approached "3.0" from a marketing angle.

My current understanding is that GNOME major releases are differentiated
primarily by binary compatibility.  The assurance is that any software
written for any major version of GNOME will work for any minor version
in that series.  While this scheme made sense to explain the transition
from GNOME 1.0 to 2.0, it is no longer workable.  There is too much
software written for 2.0 to throw it away wholesale, but more
importantly there's no overriding reason to change the desktop
infrastructure that GNOME already has.  As explained by Miguel and
others, moving to 3.0 implies a seismic shift in desktop technology away
from WIMP.  Because of this, we've been told not to expect a 3.0 release
at all.

I think this logic is fundamentally flawed.  Version numbering schemes
are all arbitrary, therefore refusing to increment versions for a
technical reason is also arbitrary.  If, by the technical definition,
GNOME will never get to version 3, then it's time to change the
numbering scheme.

Why is it important to reach new versions?  Because version numbers are
critical in marketing.  Microsoft very specifically named their
next-generation console "Xbox 360" when they could have easily called it
Xbox 2.  They wanted to avoid being a "2" while Sony charges ahead with
a Playstation 3.  Similarly, GNOME should catch up to KDEs numbers.
There's no shame in playing catchup when all you're doing is renumbering
your product. It's a simple an easy way to make yourself more comparable
with the competition.

Besides helping improve image against competition, big version numbers
get big press.  Version numbers imply progress, improvement, and new
levels in features.  Users will get excited, jaded developers will take
another look, and the faithful will be reinvigorated.  

Lastly, reaching a new big version number will avoid the current
numbering mess, where "2.10" is greater than "2.8".  Sure, once it's
explained it makes sense, but it's not intuitively obvious and makes
progress look slow.

For these reasons I think GNOME should call _something_ 3.0.  The
question is, what do you call 3.0 and when?  Hopefully, soon.  3.0 is a
great version number to be at.  3.0 is the version at which Microsoft
products become good.  3G is what people look for in their mobile phone.
More generally, three is a great number.  Just look at tripleverb.com
and see how popular 3 is for marketing.  GNOME 3.0 could utilize a
coordinated campaign for GNOME 3.0 that exploits this notion of 3s.
Therefore all we need for 3.0 is Three Big Features. 

In the end, it doesn't really matter what the Three Big Features are as
long as they are extremely user-visible.  If we're reaching 3.0 because
of marketing, it should be obvious what has changed between 2.0 and 3.0.
If I think back to the transition from 1.0 to 2.0, I don't remember much
about libraries or APIs, I remember "smooth fonts."  To me, the
transition from gtk 1.0 to 2.0 was the real Big Feature in that
upgrade.  

So Big Feature number one should be pervasive use of Cairo, composite,
and xdamage.  Gnome 3.0 should have drop shadows, fading menus, slightly
transparent notification bubbles, window animations, workspace switching
animations, and more.  It doesn't have to be gaudy, and it could be
turned off, but users have to be able to _see_ a difference, and it
should look fun.  For every GNOME application, ask "can we add any
eye-candy here?"  People like Linus will scream and moan, and that's
exactly what we want.

Big Feature number two could be multimedia support through gstreamer.
Sure, gstreamer is already on the desktop, but we can push it more.
Show off Diva or Pitivi.  Advertise out-of-the-box mp3 as well as other
formats.

Big Feature number three could be desktop search and tagging.  Push
Beagle out there.  Apple made a lot out of spotlight and we can do the
same with beagle.  Polish the nautilus tagging extension, and connect it
to beagle.  F-spot already does tagging, so that work is already done.
Tagging and Search are big web 2.0 ideas, so we could pitch GNOME 3.0 as
"Bringing web 2.0 to the desktop."

I'm not deeply immersed in the GNOME developing world, so my specific
suggestions for features may be unrealistic.  Those who work at Novell,
Redhat, and elsewhere are probably in a better position to decide
exactly what features are ready for primetime.  But I think the GNOME
community really needs a 3.0.  I think a lot of people are getting a
little depressed and weary -- although that may just be because it's
winter in this hemisphere.  Although GNOME has been making excellent
progress, it has been incremental and gradual.  Having a 3.0 goal will
give the community something to aim for and get excited about.

Thanks,
Owen Williams