[IAEP] [SLOBS] Unmaintained Projects and Teams

[Bernie Innocenti]

On Fri, 2010-07-02 at 10:06 +0200, Tomeu Vizoso wrote:

> > Proposal: draw a list of teams that should be hidden in some "Disbanded
> > Teams" page. Or, to put it positively, "Teams Awaiting Coordinators".
> > At the top of the team page, we could add a template saying "This team
> > is looking for a coordinator, ask inside". How does it sound?
> 
> Sounds great! Everything that gives more visibility to the resourcing
> problem is of great importance to SLs today.

Ok, which teams are currently inactive? These are those I never hear
anything about:

 * Education Team
 * Deployment Team (briefly revived by rgs one month ago)
 * Documentation (though dfarning seems to be up to something)

All the other teams appear to have at least some activity, so the
situation isn't as bad as it seemed initially.


> Sure not, but we need to take into account that we may not be able to
> highlight 60 projects, a more realistic number would be 6? But I leave
> this subject for our community team ;)

Good point. 6 seems like a good number. When we get there, we could do
like Apache and Eclipse do: split them between primary/secondary or
front-page/incubator projects. It could be done with an objective
criteria or by vote of the board.


> > Red Hat and Canonical got into the habit of showcasing their top
> > projects on freedesktop.org, gnome.org, kernel.org and similar umbrella
> > organizations.
> 
> Not sure I get this. You mean that they sponsor projects in those
> upstream orgs and capitalize it in terms of marketing?

No, I meant to say that RH and Canonical, who have perfectly good
project hosting sites of their own, choose to host many high-profile
projects on upstream, vendor-neutral sites.

In other words, the parallel I was trying to make is:

 Red Hat / Canonical => freedesktop.org, gnome.org...
 OLPC / SoaS => sugarlabs.org, fedoraproject.org...


> I guess we need to find the best balance for us. Organizations whose
> primary goal is supporting an upstream project use to have very clear
> that they shouldn't get into the downstream waters (no pun intended).
> I'm thinking of GNOME, KDE, LXDE, etc.
> 
> Some links for comparison:
> 
> * http://lxde.org/download - points to Debian images (but without
> making it too explicit)
> 
> * http://www.kde.org/download/ - links to distros carrying KDE
> 
> * http://library.gnome.org/misc/release-notes/2.30/#rninstallation -
> promotes "GNOME Live Media" for trying it out, points to distros for
> actual usage. GNOME Live Media is based on Foresight Linux but it's
> not explicitly mentioned.
>
> What I conclude from that is that upstream projects understand the
> need to make as easy as possible for people to try out their stuff,
> which leads to some contributors of both that upstream and of some
> distro to make customized images. Though I don't see no upstream
> project taking a distro and optimizing it for a specific piece of
> hardware, I guess Sugar's situation is special enough in this aspect.
> For SLs' is critical that our software runs as well as possible on
> OLPC machines.

For package distributions, small projects often start by making their
own binaries and then let the various distros take over.

We're doing this too, at the package level: we release tarballs and let
Fedora, Debian, OpenSUSE and Mandriva package them up autonomously.
Ubuntu is getting there too, with some help.

OTOH, to make Sugar useful in the classroom, especially in 1-to-1
computing environments, we absolutely also need customized distros. This
does not seem to be happening without our direct push. At least, not
yet.

The day OLPC ships the latest version of Sugar, I'll be glad to step
asid and move to do something else. SoaS is done by people who are both
in the Fedora and in the Sugar camp. Would it happen entirely within
Fedora? I doubt it.

Perhaps because of immaturity of the Sugar ecosystem, the usual thing
that happened to KDE and GNOME has not yet happened to us. Perhaps, it's
just that low-age education has special demands and cannot be fit into
generic mainstream distros.


> Wonder when someone will start a LTSP-ready spin for schools with computer labs.

Ask David Van Assche, he's been working on it for a while. With Sugar,
too.


> Sure, I'm in favour of trimming out the non-essential goals and
> looking at similar organizations for models. But I also see the point
> that it may be in our interest to nurture the broadest interests
> around Sugar. Crazy idea: what about an incubation lab for projects
> that don't fit 100% and that haven't grown yet to have their own
> organization?

I'm all for it! We already have all the needed pieces of infrastructure
to brew activities: gitorious, wiki, aslo. Trac and Pootle require
sysadmin action.

For projects other than activities, we offer free shell accounts with
file distribution on http://people.sugarlabs.org . Admittedly, it's a
little weak.

To improve the situation, we tried to make a deal with Launchpad and
even install our own instance of Launchpad. Project "forges" are complex
stuff, not sure we want to get there. A newer Gitorious would improve
the hosting quality without costing us anything.

-- 
   // Bernie Innocenti - http://codewiz.org/
 \X/  Sugar Labs       - http://sugarlabs.org/