[SoaS] [IAEP][DP] Announcing the creation of a SoaS Decision Panel

[Wade Brainerd]

Hi Martin,

On Tue, Sep 29, 2009 at 4:33 PM, Martin Dengler <martin at martindengler.com>wrote:

> On Tue, Sep 29, 2009 at 04:02:50PM -0400, Wade Brainerd wrote:
> > Anyway, why are we talking about the effort to maintain a distro?
>
> Because we keep getting requests to look like a distro vendor:
>
> - http://lists.sugarlabs.org/archive/sugar-devel/2009-July/016566.html
>
> - http://wiki.sugarlabs.org/go/Sugar_on_a_Stick/TODO (5 of 6 TODO items
> are distro-vendor stuff)
>
> - http://sugaronastick.com/faq/ (I'm not affiliated with this site)
>
> - endless discussions about disk layouts, e.g.
>  http://lists.sugarlabs.org/archive/marketing/2009-August/001937.html
>  http://lists.sugarlabs.org/archive/iaep/2009-August/007650.html
>
> - endless discussions about where to file bugs
>  http://lists.sugarlabs.org/archive/iaep/2009-August/007839.html
>
> I could go on.
>

Would switching to a Fedora Sugar spin distribution answer any of these
questions?  SoaS just takes Fedora and makes it a bit simpler, and gives us
the control to actually fix stuff for deployments.

If we turned all that work over to Fedora, we would just lose what little
ability we have to support deployments now.

> SoaS obviously benefits from all of Fedora's work - just look how
> > quickly it's moved from f10 to f11 to f12.  It's just a tweaked .ks
> > file and a brand name.
>
> Have you tried building it lately? It's not rocket science but it's
> not something I'd want to talk my parents through.
>

Sure, I made some tweaks to the scripts to improve the way virtual machine
images were being generated.  Like you said it's not the easiest thing in
the world, but certainly something that a school's IT staff should be able
to handle.


> > >> And, we are marketing it with success (cf. worldwide tech press
> > > >> coverage, the BBC, etc).
> > > >
> > > > How many people are using it?  Satisfied with SoaS as a distro?
> > > > What's the target deployment size, and what SL support will be
> > > > required?
> > >
> >
> > I see emails from the GPA deployment going by all the time.  Are there
> any
> > deployments using Sugar on a standard distro?  Are there often mainstream
> > media articles announcing "package foo available in latest fedora"?
>
> I hoped you would apply your example to my questions:
>
> How many people are using it?
>  GPA - 30-ish (I have no idea, really)
>
> Satisfied
>  No
>
> Target deployment size
>  30?
>
> SL support required
>  Tons of emails to IAEP and sugar-devel, Walter, new partition layout
>  proposals, new hardware support requests...
>
> How does that one-micro-deployment resource load foreshadow the load
> when SoaS becomes what you want SoaS to become?  How can that be
> achieved with the current SoaS/SL team?
>

Do we want children to use Sugar? If so, we need to do what it takes to make
that possible.  All those emails and nasty issues need to be answered.

Right now, we are the only people who can do so are on these lists, and SoaS
is the place they get fixed.

If in the future, Fedora solves all these issues for us, our kickstart file
will get a lot simpler.

> To me, SoaS has always been the answer to the question "how do I run Sugar
> > now that OLPC isn't providing OS builds?".  It works well, it's actively
> > developed, it brings in good press, so why are people questioning
> > its value?
>
> I don't see how you think I'm questioning SoaS's value.  As a
> contributor alone that would be an odd position.  I'm questioning the
> manpower and focus SL wants to bring to bear on it.  I'd like a
> statement I can point to when people ask for stuff that's out of
> scope, because the email/bug <-> commit ratio is far greater than one
> now :).
>

Lacking a better solution, I don't think we have a choice but to dedicate
the resources we have to it and try to grow the support infrastructure.  The
more schools use it, the more developers we'll attract.

I'm trying to point out that SL had better have an idea of what it
> wants out of SoaS if it wants to invest time in it.  So far it seems
> like SL wants to put in (great) marketing and get out a linux distro
> that beats Fedora in hardware and community support.


I don't understand this point, which seems to be the crux of the issue...
SoaS *is* Fedora.  Are there any circumstances where Fedora would work
better than SoaS on target hardware?

-Wade
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.sugarlabs.org/archive/soas/attachments/20090929/913f62f0/attachment-0001.htm