dfarning at gmail.com
Wed May 28 02:55:03 CEST 2008
On Tue, 2008-05-27 at 18:25 -0400, Walter Bender wrote:
> Good question to which there is not a definitive answer yet. The model
> I have been kicking around in my head is to have a small team that
> keeps its focus on top the various infrastructure needs of the
> community and raises money to support community gatherings and such
> incidentals as the filing of trademarks (expensive), etc.
> We've also been discussing other needs and models for supporting Sugar
> development and Sugar deployments. To what extent should we strive
> towards having an in-house team dedicated to such efforts? I lean
> towards a minimal footprint in keeping with the spirit of maintaining
> a diverse and distributed project, but it has been pointed out that
> model is asking perhaps too much at times. Plus it is a very young
> effort and will need some nurturing to reach a level of stability.
Is sugar labs going to support deployments? Or should the deployments be
handled by by distributions?
> We will need so some commitment of engineering resources from industry
> and other parties interested in Sugar as well as some commitment to
> Sugar Labs itself.
> These commitments would scale depending upon how much work is required
> (for a port or some necessary customization). At a minimum we'll need
> the commitment of liaisons from industry and deployments and enough of
> a community with whom they can reliably interact.
> The types of things that need to be worked on (by someone) include
> support for different distributions (and operating systems?), hardware
> platforms, localization, maintenance of existing activities, support
> for new activities, QA, documentation, evaluation, storytelling, etc.
> Some of these things require bootstrapping; some may require dedicated
We need to grow the community! All of these things can be done by
community members. A good community is like a good financial investment,
once the initial made they grow by themselves.
Redhat, encountered one of the most dangerous pitfalls when launching
Fedora. Redhat created, or at least allowed, the perception that
decisions were being make in house and the community was just there to
save on labor costs. This is a very easy trap to fall into. I can be
as innocently talking in the hallway and coming to an agreement. Thus,
it can appear to the community that decisions are made in secret. Once
Redhat created the culture on the fedora side of the house, that even
decision among friends should happen in public. The community started
Ubuntu seemed to figure this out early. I believe that is why all paid
developers at Ubuntu telecommute. All decisions are made on mailing
lists of public IRC. As a result Ubuntu has one of the most loyal
> If we leave things entirely up to hardware vendors and their partners,
> this would require an unrealistic commitment of engineering resources
> on their side (at least initially) and there is little evidence of
> their commitment to resources beyond engineering; OLPC has made such a
> commitment in the past, but it is not yet clear they will continue or
> that others would (could) follow their example.
I know I am not paying the bills at the Lab:) But, I would shy away from
the commitment angle. The goal is to create a product that
organizations have a vested interested in improving. Would Intel or AMD
be interested in improving Sugar in order to sell laptops based on their
their chips? How about Asus looking to open a new market. Would any of
the major distributions be interested in adding a Sugar based product to
their line? How about the good will angle which organization? Who
would benifit from being a proud sponsor of sugar?
> Should we choose to support just a single distribution, we are going
> to run into distribution wars both on the community and on the
> deployment side, so we really need to be at a cross-distribution
> level, which is where we are heading, but this is a lot to ask of an
> all volunteer community.
Are there any single distributions willing to support Sugar Labs until
the Lab becomes self supporting? On the other hand, being distribution
agnostic, such as gnome, KDE, Eclipse, and Openoffice seems more
beneficial. It would actually be less work to be distribution agnostic.
Every six months Sugar Labs ships a tarball consisting of the Sugar
stack. The individual distributions would package up our tarball and
put them on top of their own stacks.
> I can imagine there would be a need for Sugar consultants--both
> technical and pedagogical--but it is not clear that Sugar Labs needs
> to be more than a clearinghouse for such services.
Have you talk to Mark Shuttleworth about how the business model for
Canonical is working? When I was in the Ubuntu community a few years
ago he was very approachable. I think he would be willing to talk
because Subuntu (a sugar based ubuntu) would round out Canonical's
More information about the Its.an.education.project