[Server-devel] stability of XS 0.5

[Sameer Verma]

On Thu, Dec 18, 2008 at 3:42 AM, Martin Langhoff
<martin.langhoff at gmail.com> wrote:
> On Thu, Dec 18, 2008 at 4:48 AM, Sameer Verma <sverma at sfsu.edu> wrote:
>> Given that we still have issues cropping up with XS 0.5, are we still
>> going to call it "stable"?
>
> It's great news that you care about this. Do you have a spare
> (standard modernish x86) you can use to join the testing efforts?

Sure! I have a Fujitsu P2120 (Transmeta Crusoe ~900MHz proc, 384MB
RAM) that I've been using for testing. We use it at all the OLPC-SF
meetings. XS 0.4 works fine on it "right out of the box", but no such
love with 0.5, hence the concern. I've been following your thread on
the built-in wireless card confusion and it looks like I may have the
same issue.

>  It's
> much easier than with the XO, which requires special HW. If you get
> involved, I can organise shipping additional hw that'll be useful.

We can always dig up a relatively modern P4 for testing. What
additional hw would this be?

>
> I'll soon push a new 0.5.1 "candidate" build. You'll be CC'd so you
> can lend a hand.
>

OK. I'll wait for it.

> OTOH, all the reported issues have been diagnosed, and all but one
> fixed. Not bad for a milestone release in what is a fast development
> schedule towards 1.0.

Don't get me wrong. I really appreciate your efforts. On my recent
trip to India (http://opensource.sfsu.edu/node/593), I realized the
immense importance of the school server in environs with no Internet
backhaul. Khairat, India's first pilot, has a server, but its an old
build (160 or 161, I think) and they recently lost their backhaul, so
they don't use it anymore. The teacher asked for a lot of things that
could be fixed easily with the current XS feature set. In other
schools where they might consider getting XOs, the immense cost of
backhaul will kill the effort before it gets off the ground. The XS
fills that gap. This reminds me: We should explore sneakernet-like
e-mail at some point, but I digress.

The offline nature of content and software (wikislice, moodle, etc.)
is going to be instrumental in getting XOs to poorer school districts
that cannot afford backhaul connectivity.. In fact, I think the role
of the school server is undersold and overshadowed by the XO's
capabilities and eye candy.

I'm just concerned that folks like Manoocher
(http://lists.laptop.org/pipermail/grassroots/2008-December/000953.html)
might get misled by the "stable" monkier only to find out that the 0.5
ISO has issues.

cheers,
Sameer
-- 
Dr. Sameer Verma, Ph.D.
Associate Professor of Information Systems
San Francisco State University
San Francisco CA 94132 USA
http://verma.sfsu.edu/
http://opensource.sfsu.edu/