Re: Generating excitement in GNOME

["John (J5) Palmieri" <johnp redhat com>]

On Wed, 2007-04-25 at 22:14 +0300, Lucas Rocha wrote:
> Hei John!
> 
> Some questions:
> 
> - How is this Sneak Peek thing different from what we have now? We can
> always develop experimental things and get feedback from community as
> the things move on. If they are relevant and get enough support, they
> will naturally get into GNOME I guess. Of course there are exceptions.
> :-P

visibility.  See some of my other replys.

> - Is it just a visibility boost for those "cool" modules?

Ya.  People said nothing is happening in GNOME because they only see big
picture stuff.  This excites people into contributing (how many times
have I seen people give up on projects because they said there is no
point if it isn't part of GNOME).   It also pulls all those experimental
stuff under one roof which means more cross pollination of ideas.

> - What would change in practice for the maintainers of the "Sneak Peek" modules?

I think there would be some responsibilities.  What they are we are
trying to figure out.  One thing would be regular updates during or just
after major releases and a willingness to take input from the GNOME
community.

For instance we said absolutely no way is NetworkManager becoming an
accepted API for GNOME until it is cleaned up (it is fine for nm-applet
to use but not an API we support).  Dan was very receptive to that idea
and we eventually want to get NetworkManager as a supported API for apps
to use.  That was an easy discussion because Dan sits right next to me
so we already had that line of communication open.  But that is not
always the case.  This opens that line of communication instead of a
developer popping out of the blue asking us to add his or her project to
the next release and we end up rejecting it because it is just not
ready.  Some developers feel disappointed and we never hear from that
project again.  It is a shock for all parties involved as there were no
lines of communications open before hand. 

> - Is there anything blocking them to have better integration with GNOME?

Just that projects developed in a vacuum are developed in isolation.
That is not good for integration. 

> I like the general idea of bringing more (structured) visibility to
> those "cool stuff" that have been flowing around[1] GNOME but I think
> I didn't get the point of your idea. Also, some of those softwares
> overlap in functionality so it will be hard to provide something
> usable or consistent from this suite (is it proposed as a suite btw?
> or something like lab projects?)

More like having a central list of these project and explanations.
People can try out the best of the breed in jhbuild or garnome.  D-Bus
(a perennial example I like to use) overlapped with orbit and DCOP.  If
we didn't get the visibility we got, no one would have tried things out.
Of course there were periods of pain where people wanted to rely on it
before it was ready and others had a hard time just getting it to work
but the feedback we got was invaluable and now were are at a point were
no one really remembers those days of changing API and me telling them
tough luck ;)  Ok, so not all projects are going to be like that but it
will give structure to be a catalyst for it. 

> As an additional (and necessary) action, what I would like to see is
> objective, practical and focused groups proposing ways/plans/roadmaps
> to integrate those things in GNOME. If we don't do this, those cool
> things will stay in this parallel world forever.

Yes, I think that would be one of the requirements.  To post revised
roadmaps every so often.

> 2007/4/25, John (J5) Palmieri <johnp redhat com>:
> > Recent blog posts I did:
> >
> > http://www.j5live.com/?p=355
> > http://www.j5live.com/?p=356
> >
> > gave some people the impression that I was against excitement in GNOME.
> > That sentiment could not be further from the truth.  In fact I see and
> > participate in exciting project every day.  However being a mature
> > project we need to weigh the new and exciting with the need to support
> > stability and time tested technology.
> >
> > The exciting stuff is happening all around us it is just that it is not
> > publicized and much of it takes a long term view that does not fit well
> > into our time based releases.  It is correct that they stay this way so
> > that they can move along, experimenting without the heavy constraints
> > that becoming part of release puts on them.
> >
> > After talking with the release team to see if there are any objections,
> > as it is the release team that will be fielding the work that this
> > proposal will generate, we now open it up for comment to the rest of the
> > GNOME developer community.  We wish to start a new "Sneak Peek" release
> > module.  This module will contain API's and applications which are in
> > the process of being developed for a future release of GNOME but are not
> > yet API stable or in a form that is full acceptable for the GNOME
> > release process.  Examples may be a fully integrated Network Manager,
> > Telepathy, Gimme, Big Board, Beagle, and Tracker.
> >
> > The current proposal for rules are that the project is in a usable
> > state, seeing active development, moving towards a time based release,
> > and having a dogfoodable upstream repository.  On top of that project
> > members should be open to advice from the GNOME community and recognize
> > that being part of the module does not grantee inclusion in GNOME at
> > some later date.
> >
> > This gives project the room to experiment while letting other developers
> > keep an eye on the future directions GNOME may go in.  It is my
> > prediction, seeing the way this sort of ecosystem bloomed with
> > freedesktop modules and GNOME, that there will be early adopters who
> > will port their apps to work with the newer technologies even before
> > API's are frozen and projects make it into the GNOME releases.  This
> > will allow the projects themselves to gain real world feedback and allow
> > GNOME to move faster by having applications already utilizing the
> > technologies before they are accepted.
> >
> > This proposal is now open to comments and adjustments.
> >
> > --
> > John (J5) Palmieri <johnp redhat com>
> >
> > _______________________________________________
> > desktop-devel-list mailing list
> > desktop-devel-list gnome org
> > http://mail.gnome.org/mailman/listinfo/desktop-devel-list
> >