[Soc-coordination] Feedback / Suggestion / Discussion topic
erich at debian.org
Fri Sep 29 00:18:06 UTC 2006
So currently aj and I are scheduled to go to the Summit. I still have
very little information on the summit (only the Debian group
administrators are currently subscribed to the relevant group, not all
mentors, but there is a publicly accessible wiki).
Apparently, Google has reserved like 11 rooms for the summit's work
sessions, scheduled in 30 minute slots (but where sessions can take more
than one slot).
There are just a few generic draft ideas listed so far.
I'd really appreciate your feedback about the summer of code, how it
went for you, and how you think it could be improved. I guess you may
have written down some information already in the questionaire by
Google, but again I don't have access to them.
I wonder whether we should propose a session on some of the effects of
sponsoring Debian has recently seen. Something like "when sponsoring is
too much" or so. We could also call it
"to-dunc-tank-or-not-to-dunc-tank-tank", of course. I know that it's
somewhat odd to discuss such a thing outside of Debian, but I don't want
to talk about Debian, but bring up the issue that some people are
unhappy when others get paid.
Of course this discussion would make more sense when joey would go to
the summit... but still I would like to raise the issue somehow, and
point out what happened with Debian (mostly by quoting the latest Debian
Maybe to point out that Google Summer of Code is doing a good thing by
paying "external" students to get them involved with the projects, and
not paying e.g. "insiders" to do something. And that all the extars
mentors get is a free t-shirt. And eventually a flight to Mountain
View... I was afraid that a similar discussion to dunc tank might arise
here, but apparently few people were keen on going on that trip.
The other big issue that has come to my mind is cross-organization
projects. That is projects that "should" involve more than one mentoring
Especially with Linux distributions this should be fairly common, but I
remember having read about other cases where this also arises; mostly
But for example the "bootup time improvements" project of Debian can
also provide some useful information for other projects. I think (havn't
investigated it more closely) it's main results were the identification
of hotspots and measurement of improvement using several approaches;
these can probably be written down on a couple of pages and can then be
used by other distributions.
But other projects might be of use for multiple distributions while
consisting of a significant amount of code or integration work, that
could be shared among distributions. I'm talking less about having the
student to actually work on two distributions, but it could be good to
have more than one distribution involved during the planning phase of a
The BitTorrent protocol modification for distributing Linux packages
comes to mind, among others, and translation coordination. Or fedoras
Some things that come to my mind for maybe next year: a management tool
for chroot/dchroot/schroot/xen/uml/vmware images, whatever, on a
filesystem level. So the latter three mostly when they are run via nfs.
People might want an easy way to update their chroots/xen instances
along with their main system via the package manager, instead of having
to log on into the (possibly shutdown) systems.
Or some computer farm (I'm not talking clustery, but non-homogenous
computer farms) management toys. Like cfengine and such. These should be
able to handle multiple linux distributions; however some form of
integration with the host OS might be useful, and could be worked on
during a SoC project.
I'm looking forward to your feedback. Thanks.
erich@(vitavonni.de|debian.org) -- GPG Key ID: 4B3A135C (o_
To be trusted is a greater complement than to be loved. //\
Die kürzeste Verbindung zwischen zwei Menschen ist ein Lächeln. V_/_
More information about the Soc-coordination