Community facilitation/2009-01-21
IRC chat on #olpc (unplanned)
Jan 21 11:16:14 <greebo> hey all, ready for impromptu chat about
stuff :)
Jan 21 11:16:58 <mchua> $greeting, $name! $pleasantry, what do you $verb
$preprosition ($adjective) $noun?
Jan 21 11:17:34 <greebo> it'd be good to chat about building the
community and community-oriented process for release management of the
varioius software aspects of the olpc vision, as obviously latest news
will seriously interrupt everything :(
Jan 21 11:17:55 <m_stone> hello again.
Jan 21 11:18:25 <mchua> So, deployments. There seems to be a general
consensus that deployments are "the point" of OLPC and thus should be
given more focus, and that their needs should drive development and
support and such upstream, rather than the "push" process that's mostly
happened so far to get the tech groundwork underway.
Jan 21 11:18:37 <mchua> (how accurate does that sound? I haven't tried
summarizing this before.)
Jan 21 11:18:54 greebo GeroldKa GlibReaper gnu{- GoatCheezWork gregdek
Jan 21 11:19:09 <mchua> greebo: I think a release management discussion
is certainly within the scope of this channel. ;)
Jan 21 11:19:22 <greebo> m_stone, hi again, also sorry to hear
about your position, so annoying that so many awesome people have recently
been let go
Jan 21 11:19:23 greebo GeroldKa GlibReaper gnu{- GoatCheezWork gregdek
Jan 21 11:19:23 greebo GeroldKa GlibReaper gnu{- GoatCheezWork gregdek
Jan 21 11:19:34 <m_stone> greebo: I prefer that release-mgmt
continue to occur publicly since that was one of the criteria I laid
out for judging release managers. :)
Jan 21 11:19:47 <m_stone> greebo: thanks. it happens. :)
Jan 21 11:20:01 <mchua> especially if m_stone is around (gregdek, cjb,
and others might also have insight on how the migration to f11 is going,
which is another release cycle related discussion.)
Jan 21 11:20:02 <greebo> m_stone, totally :) awesome <- re
public process
Jan 21 11:20:15 * mchua types too slowly these days
Jan 21 11:20:59 <mchua> (warning: my laptop battery has 20 min remaining,
and this talk has >20min remaining, so I might ghost out)
Jan 21 11:21:12 <mchua> greebo: so, those
links I promised to shoot you the other day:
http://wiki.laptop.org/index.php?title=Software_discussion_2009-01-09
Jan 21 11:21:32 * vvinet has quit (Read error: 110 (Connection
timed out))
Jan 21 11:21:50 <m_stone> (for reference -- check out
http://wiki.laptop.org/go/Plan_of_Record-2008 especially the second half)
Jan 21 11:22:36 <mchua> greebo:
Jan 21 11:22:37 <mchua> http://gregdek.livejournal.com/43698.html
Jan 21 11:22:45 <mchua> GAH why does my laptop not like my copypasting.
Jan 21 11:23:34 <greebo> m_stone, looking now
Jan 21 11:23:51 <mchua> m_stone: do you know anything about the release
management process edmcnierney is using this time around? it seems to
be quite different from yours (more lightweight, etc - possibly because
8.2.1 is an interim rather than a major release.)
Jan 21 11:24:57 * befana has quit (Read error: 110 (Connection
timed out))
Jan 21 11:25:04 <mchua> (what I'm mostly trying to say is that we don't
necessarily have a culturally agreed-upon "OLPC release process," but
rather "m_stone did this for 8.2.0," - I'm not sure how widely the notion
of "this is how we do stuff" has been disseminated.)
Jan 21 11:25:44 <greebo> m_stone, what do you think of the
GNOME release process? It feels to me like there is a shift from doing
everything internally to drawing more from upstream, distribution style
(with of course all the little xo specifics added), so perhaps the GNOME
release process would be worth looking at?
Jan 21 11:26:04 <m_stone> mchua: my work is specifically
time-dated. it was for a single release in 2008 and was intended to be
revised afterward by the next person who came along.
Jan 21 11:26:44 <m_stone> greebo: I haven't followed gnome
development closely enough to have any sense of how their release-work
was carried out in practice.
Jan 21 11:26:49 <m_stone> (I've been impressed by the results
though.)
Jan 21 11:27:11 <mchua> m_stone: I'm trying to find a quick overview of
how GNOME does things, myself; see http://live.gnome.org/RoadMap
Jan 21 11:27:55 <greebo> m_stone, ok, because I've been thinking
about how we could actually build the real olpc community up, in a
self-managed and directed way, and the fact that projects like GNOME
have awesome processes and ultimately are quite sustainable so we could
learn from and imitate those
Jan 21 11:28:32 <greebo> also been thinking about all the
refugees from OLPC, who believe in the vision of one laptop per child,
but don't quite fit Sugar Labs (such as deployers, hardware, xo images,
implementation support)
Jan 21 11:28:38 * befana (n=befana@190.144.127.34) has joined #olpc
Jan 21 11:29:23 * phil_praxis has quit ()
Jan 21 11:29:44 <greebo> and how great it would be to establish a
self-directed community group that of course would collaborate with OLPC,
would no doubt work closely with Sugar Labs, and with existing groups
like OLPC Nepal and such, but would pull together the global community
so we can have sustainable software releases, community/peer support
mechanisms, etc
Jan 21 11:30:09 <mchua> we're very much not starting from scratch here;
there's already the s/olpc-release-cycle/f11 transition, support-gang,
local chapters, etc. (they're just somewhat scattered)
Jan 21 11:30:21 <greebo> totally!
Jan 21 11:30:47 <m_stone> mchua: those are conditions which make
it possible for a release process to come into existence.
Jan 21 11:31:05 <m_stone> they don't actually imply that a credible
process already exists.
Jan 21 11:31:09 * mchua nods
Jan 21 11:32:09 <m_stone> greebo: can we rewind a little bit?
Jan 21 11:32:15 <mchua> I want to make sure we're not reinventing the
wheel or forking the solution to a problem before we're sure it can't
and won't fit the stuff we want, though
Jan 21 11:32:18 <m_stone> I'm actually a bit confused about what
questions we're trying to address here.
Jan 21 11:32:33 <greebo> m_stone,
http://live.gnome.org/RoadMap/Process is more useful
Jan 21 11:32:41 greebo GeroldKa GlibReaper gnu{- GoatCheezWork gregdek
Jan 21 11:32:43 <greebo> m_stone, yeah sure, rewind away :)
Jan 21 11:32:46 <mchua> ah, thanks greebo - that's much more what I was
looking for. :)
Jan 21 11:32:48 <m_stone> so far we seem to be mainly swapping
links. :)
Jan 21 11:32:58 <greebo> I wasn't swapping links to start :)
Jan 21 11:33:03 <greebo> so I'll lay out where this started for me
Jan 21 11:33:13 <m_stone> I like swapping links because I like
sharing, but I'm not sure that it's exactly what you were trying to
accomplish. :)
Jan 21 11:34:09 <m_stone> I /think/ you were asking for thoughts -
in meandering sort of way - about how to help crystallize the kind of
community that you painted for me in your word-picture at 19:27 or so.
Jan 21 11:34:14 <greebo> There is a growing situation where OLPC
HQ are not providing the appropriate direction, leadership, tech, etc,
and they haven't really fostered a strong community so there are a lot
of people left out in the wings with nowhere to go
Jan 21 11:34:35 <greebo> Sugar Labs is a fantastic project,
and has built leadership and direction for Sugar, but (hold on one sec)
Jan 21 11:35:00 <mchua> battery, 3 min laptop time - greebo, can you
send me logs of what I'm missing between now and when I pop back online /
find an outlet?
Jan 21 11:49:05 <greebo> mchua, sure
Jan 21 11:49:05 * befana has quit ("Leaving.")
Jan 21 11:49:05 * behdad
(n=behdad@CPE001217b19226-CM0012c9c84bc4.cpe.net.cable.rogers.com)
has joined #olpc
Jan 21 11:49:05 * dirakx has quit ("Leaving.")
Jan 21 11:49:05 <greebo> so., Sugar Labs has created the mechanism
and leadership for Sugar, but the problem is that an XO/XS implementation,
or even a Sugar implementation has additional factors (hardware, OS,
support) that it doesn't make sense for Sugar Labs to do, because they
would detract from what Sugar Labs is trying to achieve
Jan 21 11:49:05 <greebo> which is fundamentally the advancement
of Sugar, and why would a Sugar hacker care about troubleshooting network
drivers (for example)
Jan 21 11:49:05 <greebo> so we have people in support land,
deployment land, OS land, and many more who don't fit into Sugar Labs,
and can't/don't fit into OLPC,
Jan 21 11:49:05 * Quozl listens
Jan 21 11:49:05 <m_stone> agreed so far, with the caveat that I
think OLPC fostered a decently strong but perhaps not /self-directed/
community. (e.g. 8.2, support-gang)
Jan 21 11:49:05 <greebo> also, unfortunately OLPC HQ keep doing
dumb things that derail the community and derail deployments
Jan 21 11:49:05 <greebo> so we need some self-direction/governance
to be able to continue with deploying the vision (which at this stage
totally makes sense on the XO, but in the future, who knows)
Jan 21 11:49:05 <greebo> m_stone, yeah, support gang is awesome,
but the community leader has just been laid off
Jan 21 11:49:05 <Quozl> (risk of any support methodology is that it
becomes a collection of workarounds, rules of thumb, and arcania, and
the code doesn't get fixed).
Jan 21 11:49:05 <m_stone> greebo: in other words, somebody needs
to pick up the pieces. :)
Jan 21 11:49:05 <greebo> my idea to support gang a few weeks ago
was to get deployments (and even perhaps Sugar Labs, Walter is interested
in this idea too) to leverage a common and good infrastructure so we can
learn from each other, but present the infrastructure in a way suitable
to their community (so OLPC Nepal may have a webform, for example)
Jan 21 11:49:05 <greebo> m_stone, yeah, heh :)
Jan 21 11:49:05 <greebo> m_stone, and we need to pick up the
pieces in the next few months before all the domain knowledge and good
will is completely leached out
Jan 21 11:49:06 <greebo> so I guess I was thinking of potentially
creating "olpc friends" as a global community thing, with an elected
council, and community leaders for different projects, a place that
people could fund volunteers to work on stuff, etc
Jan 21 11:49:06 <m_stone> greebo: strongly agreed modulo figuring
out the degree to which that expertise can be sustained.
Jan 21 11:49:06 <greebo> not sure of the specifics of how it
might look, but originally we were looking at this for the region (Aus,
NZ and the Pacific) for all the reasons above (but obviously smaller scale
of scope and need) and it is clear that a community structure is a much
greater global need now, including a broader scope of projects and such
Jan 21 11:49:06 <m_stone> greebo: so have we reached the point
where your certainty about your conclusions wanes?
Jan 21 11:49:06 <m_stone> i.e. where it's not fully clear how to
proceed and where you want to try to produce a plan that other people
will sign on to?
Jan 21 11:49:06 <greebo> m_stone, I'm always open to changing my
mind on anything :) this is really the bones of an idea
Jan 21 11:49:06 * ttuttle (n=tom@MAROON.RES.CMU.EDU) has left #olpc
Jan 21 11:49:06 <greebo> i think how to proceed could be relatively
simple, and organically grown
Jan 21 11:49:06 <m_stone> oh yes, it could be.
Jan 21 11:49:06 <greebo> so we could say that support-gang for
instance, really should be a community project (which I'm sure OLPC
HQ would encourage) and we could put in place easy ways for people to
leverage the same infrastructure for mutual gain
Jan 21 11:49:06 <m_stone> sure! I was just asking whether you had
a fully-formed consensus on this or whether you're trying to build such
a thing.
Jan 21 11:49:06 <greebo> we could establish a community process
(perhaps based on GNOME) for dealing with XO/XS releases, with the idea
that we are creating the leadership, and making it easier for community
contributions
Jan 21 11:49:06 <greebo> I would like to build such a thing :)
I'm trying to get consensus ;)
Jan 21 11:49:06 <greebo> Also, looking at ideas, at needs, at
the gaps
Jan 21 11:49:06 <greebo> A great discussion last night resulted
in a few keys gaps missing in the picture at the moment:
Jan 21 11:49:06 * m_stone listens closely.
Jan 21 11:49:06 * greebo gets her notes
Jan 21 11:49:06 * ctyler has quit (Read error: 110 (Connection
timed out))
Jan 21 11:49:06 * ctyler (n=chris@global.proximity.on.ca) has
joined #olpc
Jan 21 11:49:10 <greebo> 1) implementation support, 2)
deployment knowledge transfer and consistency, 3) xo/xs maintenance, 4)
managing/facilitating community contributions
Jan 21 11:49:32 <mchua> greebo: ...wow, I think I just got the backlog
Jan 21 11:49:45 * arjunnayini
(n=arjunnay@scn-203-76.airstudent.imsa.edu) has joined #olpc
Jan 21 11:49:53 <mchua> arjunnayini: HEY!
Jan 21 11:49:57 <greebo> mchua, I didn't realise you'd gone and
come back again, sorry :)
Jan 21 11:49:59 greebo GeroldKa GlibReaper gnu{- GoatCheezWork gregdek
Jan 21 11:50:16 <mchua> greebo: yeah, my battery just died so I ran
downstairs to the... little nook outside the speeches (where joel was
yesterday)
Jan 21 11:50:30 <mchua> greebo: np, just letting you know you didn't
need to send me logs :)
Jan 21 11:50:30 <greebo> m_stone, these are the immediate gaps,
no doubt more will come up as we go, but the existing immediate gaps
warrant figuring out a serious solution
Jan 21 11:50:33 * mchua reads backlog
Jan 21 11:50:55 <m_stone> greebo: sounds about right.
Jan 21 11:51:38 <greebo> m_stone, and if we can simplify and
largely automate the xs/xo release management process such that it can
be built upon easily enough (and tinderboxed as much as possible), we
may find the release management team for the client in particular may
be able to in the short term automate and sort out xo images
Jan 21 11:51:39 <m_stone> I've been thinking a bit about 5) acting
as a neutral market-maker/facilitator for contracts/auctions between
community members (though I don't have much to show for it yet)
Jan 21 11:51:57 * nessy (n=nessy@169.222.9.224) has joined #olpc
Jan 21 11:52:21 * mchua finishes reading backlog
Jan 21 11:52:21 greebo GeroldKa GlibReaper gnu{- GoatCheezWork gregdek
Jan 21 11:52:32 <m_stone> greebo: based on what I know today,
automation seems unrealistic to me to rely on in the near future. (it's a
good thing to invest in, but I don't expect it to solve anyone's problems
anytime soon)
Jan 21 11:52:47 <greebo> and the in the medium to short term be
able to add release images, for example: debian-based sugar implementation
for the x86 architecture, or fedora-sugar image for mips (just to be
crazy) and we look at the Debian release management to scale the releases
to make it easy to implement anywhere, derivative style
Jan 21 11:52:47 <m_stone> or rather, full automation.
Jan 21 11:52:56 <m_stone> I'd certainly believe that the scripting
could get better. :)
Jan 21 11:53:12 <greebo> m_stone, yes!!! (re 5),
abso-friggin-lutely
Jan 21 11:53:14 <mchua> m_stone + 1 on automation; there's some
medium-hanging fruit on tinderbox that cjb (in his, ah, copious amounts
of free time these days) might be able to get someone started on, but
it would take some dedicated vols time and effort
Jan 21 11:54:15 <greebo> m_stone, and we can pull in people who
are interested and have that expertise from other projects too :)
Jan 21 11:54:21 <mchua> hm, I should write up my notes (from last night,
after you went to sleep, greebo) about what an RT++-like system to help
with (1) and (2) might look like.
Jan 21 11:54:31 <m_stone> greebo: so let's take 1-5 as the things
which we want to try to build a reasonable plan around over the next
few weeks.
Jan 21 11:54:44 <mchua> wiki page?
Jan 21 11:54:44 <m_stone> greebo: maybe we'll drop one, maybe
we'll add one, but that seems like a good starting list.
Jan 21 11:55:12 <greebo> m_stone, re your number 5, we could have
an olpc marketplace, Ubuntu style...
Jan 21 11:55:47 * dfarning
(n=dfarning@75-121-133-81.dyn.centurytel.net) has joined #olpc
Jan 21 11:55:54 <greebo> m_stone, and the marketplace is a low
tech, low touch way to connect people to other people
Jan 21 11:56:01 * ysun (n=ysun@nat-pool-128-94.olin.edu) has
joined #olpc
Jan 21 11:56:09 <mchua> dfarning: hey - you probably will like this convo
Jan 21 11:56:13 <m_stone> dfarning, ysun: evening!
Jan 21 11:56:20 <greebo> if we also maintain (through the deployer
meetups) some knowledge of the status of projects, then people will be
able to find and aproach for services and support
Jan 21 11:56:31 <dfarning> good evening:)
Jan 21 11:56:39 * kristianpaul (i=paul@190.102.205.161) has
joined #olpc
Jan 21 11:56:47 <ysun> m_stone: good evening to you too :)
Jan 21 11:56:50 <m_stone> greebo: yes, that's about what I was
thinking.
Jan 21 11:56:51 <mchua> ysun, dfarning, arjunnayini: m_stone and greebo
are discussing ways to move community development forward for OLPC (as
a separate entity from SL, but obviously must work quite closely with SL)
Jan 21 11:57:06 <greebo> mchua, I think there are some good and
relatively easy things we could do to make RT nicer, but let's talk
detail in another discussion :)
Jan 21 11:57:31 <mchua> greebo: aye, that's why I'll write up those
notes later. it's a minor point.
Jan 21 11:57:41 <mchua> (more an idea that got stuck in my head than
anything else)
Jan 21 11:58:19 AaronThul abostrom arjunnayini arnd asbjornit ashok
Jan 21 11:58:23 <mchua> dfarning, ysun, arjunnayini:
http://pastebin.ca/1314105 for backlog
Jan 21 11:58:27 * dsaxena_away is now known as dsaxena
Jan 21 11:58:46 <greebo> I was thinking about maybe using "olpc
friends" as a non-profit, community driven/elected/managed body within
which we have projects (such as potentially support-gang, and release
management, deployments), community leaders, structures to make it easy
for the community to participate/contribute
Jan 21 11:59:38 <mchua> we'd need to get enough people to buy into it
beforehand that that actually becomes "the olpc community body" - how
to build consensus that this is a good idea?
Jan 21 11:59:50 <greebo> m_stone, I guess what needs to happen
asap is a meetup like this, where we pull together a load of people who
care about this, and have this discussion again, with some notice for
people to participate
Jan 21 12:00:05 <greebo> mchua, heh, see above :)
Jan 21 12:00:14 * mchua is making [[Community facilitation]]
with summary notes from this...
Jan 21 12:00:24 greebo GeroldKa GlibReaper gnu{- GoatCheezWork gregdek
Jan 21 12:00:27 <mchua> greebo: nice. :)
Jan 21 12:00:38 <m_stone> greebo: I don't see it as critical that
everyone be on the same page at the same time instantly.
Jan 21 12:00:39 <greebo> if the community body establishes an
interim community council with well respected people then that would go
a long way to helping people buyin to it, and give it some credibility
Jan 21 12:00:51 <greebo> (I'm a sucker for structure, but realise
we need to figure out the right structure
Jan 21 12:00:52 <greebo> )
Jan 21 12:00:56 <m_stone> greebo: in other words, I think we can
give ourselves a few weeks to work things out so long as we're making
visible public progress each week.
Jan 21 12:01:14 <greebo> m_stone, so, not everyone, but we need
thought leaders on the same page, and we need community leaders (official
and unofficial) on the same page
Jan 21 12:01:17 <m_stone> (at least, that worked for me last
March-April) :)
Jan 21 12:01:31 <greebo> m_stone, yeah, cool, good plan :)
Jan 21 12:01:46 <m_stone> greebo: again, not instantly. we just
need them to wind up on the same page within, say, two months.
Jan 21 12:02:06 <m_stone> (that's my feeling today, anyway)
Jan 21 12:02:07 * jrb has quit (Read error: 110 (Connection
timed out))
Jan 21 12:02:12 <greebo> m_stone, btw, as a cool aside, I'm looking
at integrating puppet to the server and clients, for real management,
rollout capability, updates, and apps management. cool hey :)
Jan 21 12:02:36 <mchua> greebo, m_stone:
http://wiki.laptop.org/go/Community_facilitation is a stub, please fill
Jan 21 12:02:37 <m_stone> heh. I thought you'd appreciate my
"six-overlapping-but-not-quite-perfect" suggestions
Jan 21 12:02:38 <greebo> m_stone, you're absolutely right,
definitely, and we can lead by example with the process.
Jan 21 12:02:58 * Sargun has quit (Remote closed the connection)
Jan 21 12:03:32 <greebo> there is a group here locally who also
are discussing this, so I'll take these notes back to them too, we're
going to have a community discussion on friday afternoon (aussie time),
so we'll document our chat then too
Jan 21 12:03:38 <greebo> mchua, awesome
Jan 21 12:03:44 <mchua> greebo: the olpcfriends discussion is a place
I'd like to roll http://wiki.laptop.org/go/Board of directors community
representative into, if there's room (later down the line, once there's
more of a community to represent.)
Jan 21 12:03:57 <m_stone> greebo: sounds good. do you want to sign
off to go give them your full attention now?
Jan 21 12:04:03 <m_stone> or do you want to do some more
brainstorming here?
Jan 21 12:04:19 * mchua has already missed half a talk; is here
anyway
Jan 21 12:04:21 <m_stone> (I expect that I'll be around later this
evening as well, so feel free to go play with them now while you've got
the f2f time)
Jan 21 12:04:42 <m_stone> since we're now agreed on the basic
strategy -- (namely, come up with plausible ideas, tell everyone about
them, and if they're persuasive, merge them into The Constitution)... :)
Jan 21 12:04:46 <greebo> m_stone, happy to do more brainstorming
for a little while, this is really cool, I'm going to chat to the support
gang people about it at the next support gang meeting
Jan 21 12:04:52 <m_stone> good, good.
Jan 21 12:05:16 <m_stone> so reviewing your list of
things-to-assume-some-responsibility-for....
Jan 21 12:05:42 <m_stone> 1) implementation support, 2)
deployment knowledge transfer and consistency, 3) xo/xs maintenance, 4)
managing/facilitating community contributions, 5) neutral market-making
Jan 21 12:05:53 * mchua realizes that much of
her cynicism/hesitation over the last few days has been
possessiveness/fear-of-possessiveness, kill -9s that process
Jan 21 12:06:04 <greebo> k, I'm going to continue chatting to
people here, updating the page, and within the week blog about this idea
with a link to the page to get others thinking along the same line
Jan 21 12:06:13 <m_stone> mchua: don't sweat it; you're far from
alone in that. :)
Jan 21 12:06:17 <greebo> I'm also going to facilitate a local
discussion about this on friday, and chat to support gang
Jan 21 12:06:27 <m_stone> greebo: good. I'm happy to spam mailing
lists some more. :)
Jan 21 12:06:38 <greebo> oh, I thought you meant personal
responsibility m_stone :)
Jan 21 12:06:40 <mchua> m_stone: yeah, I've been bothered by it lately
and trying to find out how the hell to get rid of it... I think I just did
Jan 21 12:06:52 <m_stone> greebo: no, I meant collective
responsibility.
Jan 21 12:06:56 greebo GeroldKa GlibReaper gnu{- GoatCheezWork gregdek
Jan 21 12:07:10 * jrb (n=jrb@68-116-198-66.dhcp.oxfr.ma.charter.com)
has joined #olpc
Jan 21 12:07:18 <m_stone> greebo: i.e. things that we're claiming
we want to have happen in a relatively timely, transparent, predictable
manner that we don't think are going to happen on their own.
Jan 21 12:07:24 <mchua> m_stone, greebo: can we agree to use
http://wiki.laptop.org/go/Community_facilitation (or some other wikipage
location) as a hub for notes of these discussions, so there's an easy
way to update ourselves re: the current status?
Jan 21 12:07:40 <greebo> m_stone, cool, let's try to have more
personal discussions with people for the first week or two, to get buy-i
without people feeling this is a done deal, or yet another thing being
done to them
Jan 21 12:07:42 <m_stone> mchua: you can try. I doubt it will
work. :)
Jan 21 12:07:53 <m_stone> mchua: (mainly because I think we're
going to go through a lot of drafts.)
Jan 21 12:08:12 <m_stone> greebo: don't worry. it'll take more
than a week. :)
Jan 21 12:08:31 <mchua> m_stone: that's why i'd like to have those drafts
on the wiki
Jan 21 12:08:35 <m_stone> greebo: the thing is that, if we do
our jobs well, it will probably /run/ for more than a few weeks as well
before people get tired of it.
Jan 21 12:08:36 <mchua> or the logs linked to from a central page
Jan 21 12:08:39 <cjb> hola all
Jan 21 12:08:44 * arnd has quit (Remote closed the connection)
Jan 21 12:08:59 <cjb> FYI, edmcnierney_away's asked me (today) to take
over as 8.2.1 release manager.
Jan 21 12:09:01 <m_stone> mchua: yes, yes. one just has to do the
hard work of transcribing things. where do you think all the commentaries
on [[User:Mstone]] came from? :)
Jan 21 12:09:04 <mchua> m_stone: centralized communication tends to help
distributed action.
Jan 21 12:09:13 * m_stone breaks down in hysterical laughter.
Jan 21 12:09:21 <mchua> m_stone: well, transcription is one of those
old habits ofmine.
Jan 21 12:09:31 <m_stone> mchua: I'd never noticed. :)
Jan 21 12:09:40 <greebo> cjb, right, how are you feeling about
that? comfortable? drowning?
Jan 21 12:09:40 <cjb> sorry, will try to catch up quickly
Jan 21 12:10:01 <mchua> cjb: ...oh shit, I need to get you QA stuff
then. sorry I've been so quiet on that front; I've been waiting to hear
shouts of what was needed
Jan 21 12:10:05 <cjb> greebo: I'm impressed by GNOME management, but
I'm wary because I think it necessarily took ten years to scale up to
that level of coordination and shared interest.
Jan 21 12:10:10 <mchua> (but suppose I should just start plunging through
tests during my "i'm not sleeping!" time)
Jan 21 12:10:34 * CosmicPenguin has quit ("leaving")
Jan 21 12:10:36 <greebo> cjb, because it would make sense to use
either this release or the next to transition into a community oriented
process, and doing it while there is someone able to commit seroius time
at that end will hopefully reduce the pain in the transition :)
Jan 21 12:10:52 <mchua> greebo: lca is your 'hood, so you're going to
be way better at those discussions here than I will be. Let me know when
I can follow you around and quietly transcribe things from the corner. ;)
Jan 21 12:11:00 <cjb> greebo: yep, that all sounds good
Jan 21 12:11:15 <cjb> greebo: I'm a heavy advocate of the "our next
software release is called Fedora 11" plan
Jan 21 12:11:34 <mchua> I'm quite determined to make these conversations
public record - all of them, including the in-person ones, in-IRC ones,
etc... we should do view-source for the process of creating this community
as well.
Jan 21 12:11:38 <greebo> cjb, no, see the 6 month release process
was created mainly by Jeff (Waugh) and once it was created, they had a
major release and then it became quite smooth quite quickly
Jan 21 12:11:43 <cjb> (I can't decide whether it was my idea or
whether it was surreptitiously planted in my head by Ed and Gregdek,
so I'll just say that I like it ;-)
Jan 21 12:11:54 <_40oxo_> hi, i asked in #olpc-help but there is
more activity here.. does anyone know how to install the Browse activity?
Jan 21 12:12:05 * m_stone is not a heavy advocate of the "our next
software release is called Fedora 11" because he doesn't hear the F11
folks make much noise about supporting the features deployments claim
to care about...
Jan 21 12:12:08 <greebo> cjb, and the thing is we already have
(right now anyway) an interested community we can leverage, we just need
a process they can participate in
Jan 21 12:12:11 <mchua> _40oxo_: I'll go to #olpc-help and give you a
hand - I'll see you in that channel.
Jan 21 12:12:13 <m_stone> but I'm open to persuasive arguments to
the contrary.
Jan 21 12:12:13 <cjb> _40oxo_: in 8.2.0, use the activity updater
control panel
Jan 21 12:12:39 <greebo> mchua, if you can transcribe friday's
meeting, that'd be awesome!
Jan 21 12:12:50 <cjb> greebo: sure
Jan 21 12:12:54 <m_stone> (I certainly advocate doing lots of work
in Fedora though so that it can be used as a good base for getting the
things that we actually want.)
Jan 21 12:13:46 <greebo> _40oxo_, yes, put it on a usb key
(download from the Activities page on wiki.laptop.org) and then put the
usb key into your xo (I assume you have an xo)
Jan 21 12:13:47 <m_stone> ((though, on the side, I'd also like to
see a credible competitor))
Jan 21 12:13:59 <_40oxo_> cjb: i have build 767, i think that is
8.2.0? the activity updater only seems to update existing applications,
maybe i'm wrong?
Jan 21 12:14:13 <greebo> then you can either click on it from
the journal, or use the olpc-install (I think) command line if you like
Jan 21 12:14:21 <cjb> _40oxo_: you're (hopefully) wrong :)
Jan 21 12:14:37 <m_stone> _40oxo_: (yes, 767 == 8.2.0)
Jan 21 12:14:49 <cjb> m_stone: yes, the point of the next release,
in my mind, is to produce a useful base.
Jan 21 12:14:57 <cjb> anything we manage to do past that is a bonus
Jan 21 12:14:59 <mchua> greebo: I transcribe *everything.*
Jan 21 12:14:59 <greebo> _40oxo_, or, you could reimage with the
application bundle if it is new
Jan 21 12:15:06 <greebo> mchua, awesome
Jan 21 12:15:11 <ysun> _40oxo_: I believe you can also download browse
onto a usb stick, open that in journal, and install it that way
Jan 21 12:16:02 <cjb> greebo: I still share your enthusiasm for how
the GNOME process works; I just think it's going to take a couple of
releases before we manage to attain a decentralized set of organizations
who each have responsibility for different parts of the stack, and so on
Jan 21 12:16:26 <_40oxo_> ysun: thank you, do i put the Browse.xo
file on the usb key? or the Browse.activity dir?
Jan 21 12:16:31 <mchua> _40oxo_: see
http://wiki.laptop.org/go/Customization_key
Jan 21 12:16:35 <cjb> _40oxo_: the .xo file
Jan 21 12:16:44 <greebo> cjb, I totally agree this will take time,
but getting the release management process itself right doesn't need to
be too overwhelming/overbearing
Jan 21 12:16:48 <_40oxo_> alright i'll try that. thanks a lot
Jan 21 12:17:00 <mchua> _40oxo_: ...although you probably don't need
that, actually, unless you want to install multiple other Activities at
the same time
Jan 21 12:17:50 <cjb> mchua: I'm glad you're in Australia and you
shouldn't feel bad about not doing more 8.2.1 work, because I haven't
been doing it either ;-) You're certainly welcome to claim ownership
of anything you're interested in claiming ownership of.
Jan 21 12:17:53 <greebo> so far as I see it, an xo image has
its upstream sources (fedora, sugar, other apps), it's special hacking
(xo specific drivers, any integration work needed) and so at each of
those levels the people repsponsibile for the different parts of the
stack already exist
Jan 21 12:17:58 <m_stone> cjb: what can I say? "achieving a useful
base" seems rather weak to me.
Jan 21 12:18:18 <cjb> greebo: So, our reason for going with Fedora 11 is
that we don't want to have any further release management process past "we
need to get our stuff to the Fedora folks before they freeze for release"
Jan 21 12:18:31 <greebo> similar to how ubuntu has debian upstream,
adds special sauce, and ubuntu is a derivative of debian. I see the xo
image as a derivative of sugar/fedora with special sauce
Jan 21 12:18:33 <mchua> m_stone: "building atop Fedora's already
well-established release process to create our own?"
Jan 21 12:18:45 <cjb> greebo: because when you have been 1 and 3
software engineers left, depending how you count them, you don't want
one of them to be a manager of anything, especially releases :)
Jan 21 12:18:59 <m_stone> cjb: so obviously OLPC shouldn't be in
the business of managing releases.
Jan 21 12:19:11 <cjb> m_stone: yes. that's what I'm saying.
Jan 21 12:19:37 <cjb> and I'm sorry you find the "put stuff into
Fedora 11" strategy weak, but I'd rather be weak than delusional.
We can't commit to anything stronger right now.
Jan 21 12:19:39 <m_stone> cjb: and that part is good and fine.
Jan 21 12:19:50 <mchua> the question, in my mind, is whether the needs
of OLPC deployments can not be met by people within those deployments
participating in SL and the fedora-olpc SIG.
Jan 21 12:19:53 <m_stone> cjb: who's the "we" you're speaking for?
Jan 21 12:20:12 * ctyler has quit (Read error: 110 (Connection
timed out))
Jan 21 12:20:15 * igor__ has quit (Client Quit)
Jan 21 12:20:18 <greebo> cjb, right, if that's the case, we could
probably encourage others to do release management if you don't want to
own it. Or did I misread the above comment? :P)
Jan 21 12:20:24 * kristianpaul has quit (Read error: 104 (Connection
reset by peer))
Jan 21 12:20:29 <cjb> m_stone: OLPC's sphere of responsibility, I guess.
Jan 21 12:20:39 * arjunnayini
(n=arjunnay@scn-203-76.airstudent.imsa.edu) has left #olpc
Jan 21 12:20:52 <mchua> if the answer is "no, they can't be met,"
then a separate release process, community around it, etc. makes total
sense. If the answer is "right now, migration up to SL and F11 and
having deployments work directly with those communities will suffice,"
the extra work isn't (yet) needed.
Jan 21 12:20:54 <m_stone> cjb: well, please stop. OLPC had its
chance to take responsibility and it blew it.
Jan 21 12:20:56 * ctyler (n=chris@global.proximity.on.ca) has
joined #olpc
Jan 21 12:21:01 <cjb> greebo: I guess you're assuming that we would
want to make a separated release if we could
Jan 21 12:21:14 <cjb> greebo: I'm not really convinced of that
Jan 21 12:21:20 <m_stone> cjb: the thing to do now is to figure
out what actually /needs/ to be done and then to find a way to make it
possible, or to change what needs to be done.
Jan 21 12:21:28 <_40oxo_> in journal, do i press "copy" to install
an activity?
Jan 21 12:21:35 <cjb> mostly because I agree with Michael; we had a
chance at being Software Overlords, and it didn't work out very well
Jan 21 12:21:42 <greebo> mchua, no, because any real deployment
will have needs that are beyond the scope of fedora and beyond the scope
of sugar, the release team is the group responsibile for meeting the
deployment needs. My idea is that part of the roadmapping process for each
release involves input from the deployment sig (for want of a better word)
Jan 21 12:21:42 * gregdek has quit (Read error: 110 (Connection
timed out))
Jan 21 12:21:44 <cjb> _40oxo_: I think you press "resume"
Jan 21 12:21:47 <cjb> on its detail page
Jan 21 12:22:06 <_40oxo_> there is no resume... there is a "start"
but that doesn't do anything
Jan 21 12:22:14 <m_stone> cjb: so I see two interesting questions
here.
Jan 21 12:22:22 <m_stone> one is: can Fedora meet the needs of
deployments?
Jan 21 12:22:23 <cjb> _40oxo_: hm, it should do something
Jan 21 12:22:35 <m_stone> with no down-stream modification required.
Jan 21 12:22:53 <ysun> _40oxo_: what happens if you just double click
on the activity? if you can get it to start at all, it'll stay in your
journal after that
Jan 21 12:23:03 <greebo> cjb, I'm not suggesting we need overlords,
I'm suggesting we need a release management team (and it should be
community run/directed with anyone interested involved including from
OLPC if that fits
Jan 21 12:23:12 <cjb> my answer is no. (but then, software rarely meets
the needs of deployments, aswe've found; this isn't very surprising.)
Jan 21 12:23:16 <mchua> m_stone: I was operating the assumption that
the answer to that question was "yes" when I started talking with greebo
last night (hence the first diagram I drew, Pia.)
Jan 21 12:23:28 <mchua> well, s/yes/close-enough.
Jan 21 12:23:32 <m_stone> mchua: do you have evidence that supports
your claim?
Jan 21 12:23:49 <greebo> m_stone, asking fedora to meet the needs
of deployments is not a sustainable process, and not reliable, it'd be
like Ubuntu asking Apache to support the needs of a particular client
Jan 21 12:24:12 <mchua> m_stone: it's my understanding that the current
"port to f11" strategy that cjb proposed the other week would give us
identical functionality to our current stable release, with the fedora
process
Jan 21 12:24:13 <cjb> greebo: it's an interesting suggestion. I guess
I'm saying that OLPC has so few resources to commit to this that my
current determination of what's best to do with them is to turn the OLPC
software into Fedora.
Jan 21 12:24:38 <m_stone> greebo: I think the implicit assumption
was that "oh, deployments will understand that in order to get new
releases, they have to become productive members of Fedora and shape
Fedora to meet their needs"
Jan 21 12:24:48 <mchua> m_stone: if that's the case, then fedora will
be at the point where it fulfills deployment needs as well as OLPC
previously had been doing (which isn't perfect, but /is/ a start) and
can move forward much more easily from there, as they have practices
and processes and people already.
Jan 21 12:24:50 <greebo> fedora should focus on making fedora
awesome, sugar focused on making sugar awesome, and our olpc community
focused on making awesome releases for the xo building on the work of
fedora and sugar, but filling the gaps with out community
Jan 21 12:25:04 <cjb> greebo: and I think the confusion here is because
m_stone and yourself thinkwe're trying to accomplish just as much as we
were doing before when we do this, but I think that we aren't.
Jan 21 12:25:13 <m_stone> greebo: i.e. the sense was that Fedora was
a sort of skin or costume which the body of people interested in XO-based
software distros could slip on in order to get something workable.
Jan 21 12:25:25 <cjb> m_stone: that sounds like a good way of putting it
Jan 21 12:25:30 <_40oxo_> ysun: double-clicking on the activity
does nothing.. and i can't start it. maybe there is something wrong with
my .xo bundle. i made it from the git sources. is there a place where i
can find all the latest .xo files, or a url i can wget? i couldn't find
Browse.xo on the olpc wiki in elinks :) hehe
Jan 21 12:25:31 <mchua> m_stone: for the time being, anyway
Jan 21 12:25:50 <m_stone> cjb: I'm actually trying to accomplish
more.
Jan 21 12:25:57 <greebo> cjb, so this is where we change our
perspective from what OLPC can do with its limited resources, and what
our expansive and enthusiastic community can do with its much larger
resources (and expertise)
Jan 21 12:26:00 <m_stone> cjb: (I just think that it's going to
take a while)
Jan 21 12:26:05 <mchua> I mean, I can see how deployment needs *will*
expand beyond the SL+fedora-olpc SIG structure that's the current plan.
Jan 21 12:26:06 <cjb> I'm welcome to hearing ideas about how we can
use our maybe one-third of a employee's time to help make good XO releases
Jan 21 12:26:09 <mchua> It's going to, inevitably.
Jan 21 12:26:11 <cjb> other ideas, I mean
Jan 21 12:26:16 <cjb> but so far this is the best I can come up with
Jan 21 12:26:25 <m_stone> cjb: okay. that's helpful to know.
Jan 21 12:26:27 <mchua> But right now, it isn't, and this is a start
towards the larger, grander plan of the olpcfriends proposal.
Jan 21 12:26:43 <m_stone> cjb: I can see why it's an attractive
solution from your standpoint. I might even advocate for the same if I
were in your shoes.
Jan 21 12:26:46 <cjb> it involves deployments taking on more work
than they had before -- more precisely, in coordinating the extra work
amongst themselves
Jan 21 12:27:08 <greebo> _40oxo_, seriously, please just reimage,
it will save you a lot of time, so long as you don't have anything on
the laptop you need to save
Jan 21 12:27:11 <mchua> i.e. I see the olpcfriends proposal as something
we want to grow into, but *not* where we are, or even what we need as
a first step, now.
Jan 21 12:27:42 <mchua> _40oxo_: If you have
a thumbdrive, reimaging instructions are here:
http://wiki.laptop.org/go/Release_Notes/8.2.0#If_your_XO_is_not_connected_to_the_internet
Jan 21 12:28:07 <cjb> I guess I worry that doing anything more than
I've proposed would send the mixed-message that OLPC still exists to
provide customized and recent software for its deployments
Jan 21 12:28:08 <_40oxo_> greebo: alright, but why is it so hard
to install an activity? why is reimaging the easiest solution?
Jan 21 12:28:17 <cjb> and I hate over-promising to people, so that's
not a message I'm comfortable giving
Jan 21 12:28:23 <greebo> cjb, cool, I understand the pressure
you're under, and the ridiculous lack of resources now available, but I
think that moving to the community run model will help you rather than
make it harder, will streamline the process and make it such that you
can be part of a release team, rather than solely respoonsible
Jan 21 12:28:33 <mchua> _40oxo_: usually, you use the Browse Activity
to install Activities.
Jan 21 12:28:37 <mchua> _40oxo_: It's quite easy then.
Jan 21 12:28:44 * m_stone thinks that cjb was an excellent member
of the 8.2 release team.
Jan 21 12:29:01 <mchua> _40oxo_: And that's one (of many) reasons
Browse is on the default XO image - I'm not sure how you got an XO
without Browse.
Jan 21 12:29:08 <cjb> greebo: that all sounds good to me :) (For
the record, I guess I think that my model is also a community-run one.
I'm not proposing centralized ownership of anything.)
Jan 21 12:29:20 <greebo> cjb, re your concern about messaging,
if the release management was publicly done through olpcfriends, then
that would fix the message
Jan 21 12:29:28 <mchua> _40oxo_: reflashing would give you an image with
Browse, and at that point you're bootstrapped into being able to install
other Activities you want.
Jan 21 12:29:29 <_40oxo_> mchua: i removed it by accident :)
Jan 21 12:29:31 <cjb> greebo: that sounds like it could work
Jan 21 12:29:41 <mchua> _40oxo_: ahh. :)
Jan 21 12:29:41 <cjb> greebo: do you think there might be room for
both approaches?
Jan 21 12:29:49 <greebo> _40oxo_, yeah, what m_stone said. That
you don't have the Browse activity already says to me that your machine
wasn't imaged correctly from the start
Jan 21 12:29:58 <cjb> something like, OLPC HQ is going to work on
making sure everything's in Fedora so that future Fedora releases are
a usable XO platform
Jan 21 12:30:10 <cjb> and olpcfriends is going to work on making
deployment-relevant releases?
Jan 21 12:30:14 <m_stone> cjb: that's something we need to find
out. :)
Jan 21 12:30:21 <cjb> much like a Fedora -> RHEL model, where
olpcfriends is RHEL
Jan 21 12:30:30 <mchua> I see cjb's current plan as step 1 towards
olpcfriends, personally.
Jan 21 12:30:33 * alemartinazzo
(n=alexandr@189-46-34-26.dsl.telesp.net.br) has joined #olpc
Jan 21 12:30:37 <greebo> cjb, yeah, but the release management
needs to itself be community driven, and fedora isn't the right
community/place for that
Jan 21 12:30:46 <cjb> greebo: the release management of what?
Jan 21 12:31:00 <mchua> i mean, the community release management would be
management of... what, right now? a fedora-based, sugar-running release.
Jan 21 12:31:03 <greebo> cjb, the release management of (at this
stage) images for the XO
Jan 21 12:31:12 <cjb> I think Fedora's certainly the right place for
Fedora's release management, which is all I proposed to use it for
Jan 21 12:31:14 <mchua> and cjb is trying to get a fedora-based,
sugar-running release to happen.
Jan 21 12:31:29 <mchua> and that's a starting point.
Jan 21 12:31:33 <dfarning> This discussion is weird - I keep reading
the answers before the questions... must be latency issues:)
Jan 21 12:31:34 <cjb> I'm saying that if olpcfriends can take that
upstream and overlay its own customization and release, that sounds like
a fine plan
Jan 21 12:31:44 <greebo> cjb, for one reason, we may find the
olpcfriends release management team may want to create an automated
ubuntu/sugar image that runs on x86, or an opensolaris image that runs
on sparc.
Jan 21 12:32:03 <cjb> greebo: why do you think I was proposing having
Fedora do olpcfriends releases?
Jan 21 12:32:07 <cjb> I think that's our miscommunication
Jan 21 12:32:25 <greebo> the release management team should be
wedded to neither OS, nor hardware platform, as I see olpcfriends as
being wedded to the vision of olpc, but not OLPC HQ. Sugar is pretty
much a given I think :)
Jan 21 12:32:30 <cjb> I want Fedora to keep doing Fedora releases,
and have them start to (they don't right now) boot on the XO usefully
and contain XO-useful software
Jan 21 12:32:38 <greebo> cjb, sorry, I wasn't suggesting that
you said that
Jan 21 12:33:03 <greebo> cjb, but you were talking about pushing
a lot of the work to fedora, and although I completely agree that fedora
updates need to be pushed upstream
Jan 21 12:33:09 <cjb> and from there, olpcfriends can create all manner
of distro-agnostic images and tools
Jan 21 12:33:30 <greebo> it isn't the right place for olpc specific
devel/management. leadership should come from the release management team
Jan 21 12:33:36 <cjb> greebo: well, I want to push it somewhere, which
probably means a specific distro, and the Fedora community is basically
offering to help
Jan 21 12:33:47 <cjb> greebo: we laid off our release management
team, sorry
Jan 21 12:34:00 <greebo> cjb, awesome, we're on the same page
Jan 21 12:34:11 <greebo> damned reading and writing IO issues :)
Jan 21 12:34:38 <m_stone> may I try to summarize?
Jan 21 12:34:47 <mchua> m_stone: please
Jan 21 12:34:48 <cjb> m_stone: that'd be superb, thanks
Jan 21 12:34:51 <greebo> cjb, we can rapidly build a release
management team with community people (like m_stone for instance) who
want to fdo this
Jan 21 12:38:34 * mchua curses her network card switch being in
an accidentally-switch-offable place
Jan 21 12:38:42 * mchua requests backlog from the last 5 mins
Jan 21 12:39:05 * Disconnected ().
**** ENDING LOGGING AT Wed Jan 21 12:39:05 2009
**** BEGIN LOGGING AT Wed Jan 21 12:39:20 2009
Jan 21 12:39:20 * Now talking on #olpc
Jan 21 12:39:20 * Topic for #olpc is: Web support:
http://forum.laptop.org | IRC support: #olpc-help | Deployment:
http://radian.org/notebook/first-deployment | GUI, activities: #sugar |
Content: #olpc-content
Jan 21 12:39:20 * Topic for #olpc set by neuralis at Mon Feb 4
21:44:05 2008
Jan 21 12:39:38 <m_stone> cjb: yes. OLPC cannot do that any more.
Jan 21 12:39:41 <m_stone> maybe olpc-friends can.
Jan 21 12:39:48 <m_stone> that's what greebo and I are trying to
figure out.
Jan 21 12:39:51 <m_stone> (or maybe Fedora can.)
Jan 21 12:39:54 <cjb> it's helpful for me to have a consistent message,
so I guess that's why the "we're not doing that any more, but we sort
of are" message is unattractive to me so far
Jan 21 12:39:55 <m_stone> (but I'm skeptical.)
Jan 21 12:40:01 <greebo> cjb, yeah, that's what we largely get
now from olpc, and we could get from the release team (doesn't matter
who provides it, so long as it is provided :)
Jan 21 12:40:11 <cjb> greebo: I'd love for that team to exist
Jan 21 12:40:21 <cjb> I don't think my Fedora proposal hurts that
team's position
Jan 21 12:40:24 <mchua> can someone shoot me the backlog starting from
m_stone's question on whether it's useful to summarize? I accidentally
switched off my card...
Jan 21 12:40:32 <cjb> it's just something I can help with to give that
team a base to work with in the future
Jan 21 12:40:44 <cjb> in short, I think the Fedora stuff is something
this olpcfriends release team would need to do anyway
Jan 21 12:40:45 <greebo> cjb, cool, I'm keen to help facilitate
creating this process, and the team/structure for it to happen. I'm not
a release manager but I know a lot of great ones and we can learn from
and apply the best practices
Jan 21 12:41:11 <cjb> I should head off momentarily, and it sounds
like greebo wants to too, so maybe a good time to break
Jan 21 12:41:18 <m_stone> true.
Jan 21 12:41:20 <mchua> I think we are looking at different points along
the same path, but want to end up in the same place.
Jan 21 12:41:24 <cjb> I think the short answer from me is "olpcfriends
+1"
Jan 21 12:41:38 <greebo> cjb, fedora people contributing is great,
but I'd suggest release management of the xo image not be a fedora
project, because that removes the vendoir/distro neutrality which which
ensure the longevity and inclusiveness of this project
Jan 21 12:41:48 <cjb> and that the best thing I can think I can do to
help olpcfriends, counter-intuitively, is go make sure OLPC's current
stuff is all in Fedora
Jan 21 12:41:56 <mchua> I think that existing communties (right now,
the Fedora-OLPC SIG + SL; possibly soon $other_distro-OLPC SIGs) are
capable of picking up much of this work, and that if we support their
ownership of that work (rather than starting by creating our own org),
then the growth of an olpcfriends-like org will more naturally occur,
as the now-more-invested-in-OLPC SIGs feel the increased need to connect
with each other.
Jan 21 12:42:01 <m_stone> cjb: yes, that's not counter-intuitive
at all.
Jan 21 12:42:05 <mchua> greebo and m_stone are just ahead of me on
this. ;)
Jan 21 12:42:11 <greebo> if the "olpc release team" have fedora
people on the team, that's fine, but it should be a team others can join,
and that maintains an open mindedness about options for future releases
Jan 21 12:42:15 <greebo> if that makes sense
Jan 21 12:42:32 <cjb> greebo: that works. I'm all about
distro-neutrality; Fedora just seems to be particularly interested in
making sure their distro is useful for us at the moment, so it makes
sense to take advantage of it.
Jan 21 12:42:51 <cjb> greebo: the "olpc release team" is a bad name,
because it implies that OLPC the company has a release team
Jan 21 12:43:01 <m_stone> cjb: (sure. though you might find that
if you gave talks at debconf, debian people would also be interested.)
Jan 21 12:43:10 <greebo> cjb, some of the stuff may not fit
in fedora, and we need to make sure is available for people to port
to other implementations (eg, an ubuntu xo distro), similar to ubuntu
having their ubuntu bits they add to debian
Jan 21 12:43:12 <cjb> but if you mean the olpcfriends (for want of a
name) release team, I totally agree
Jan 21 12:43:43 <cjb> greebo: yeah! if a country came to
OLPC-the-company and said they wanted to run DebXO on their deployment,
I'd be happy with that
Jan 21 12:43:50 <greebo> mchua, but they may not need to connect
with each other, they may just fork and then 12 months later we have
complete division on a code and vision level
Jan 21 12:44:03 <cjb> (this is actually a new development -- we've
always had a very strong connection to Red Hat in the past that led to
us strongly recommending Fedora)
Jan 21 12:44:09 <cjb> (and I think that's going to be less so in
the future)
Jan 21 12:44:26 <greebo> olpcfriends I believe will avoid the
complete fragmentation of tech and perspective which will threaten our
capacity to actually deliver this stuff to kids in a sustianable way
Jan 21 12:44:28 <m_stone> cjb: the end of a monopoly. :)
Jan 21 12:44:36 <cjb> m_stone: yup!
Jan 21 12:44:49 <cjb> greebo: all sounds good
Jan 21 12:44:54 <cjb> is there anything I can do to help in the
near future?
Jan 21 12:45:01 <cjb> I guess I shouldn't offer webspace and mailing
lists
Jan 21 12:45:04 <m_stone> cjb: sure! keep chatting with us. :)
Jan 21 12:45:05 <cjb> since the whole point is to decentralize
Jan 21 12:45:06 <greebo> cjb, yeah, we need a name, that isn't olpc
(company clash), isn't xo (future hardware potential clash), isn't sugar
(sugar labs clash) and yet implies all of these to a limited degree ;)
Jan 21 12:45:13 <cjb> haha
Jan 21 12:45:15 <cjb> awesome
Jan 21 12:45:29 <greebo> "olpcfriends release team", I like it :)
Jan 21 12:45:33 <mchua> "the child-laptop bijection project"?
Jan 21 12:45:47 <cjb> mchua: catchy!
Jan 21 12:45:51 <m_stone> mchua: ++
Jan 21 12:45:59 <cjb> greebo: do you think there are naming issues with
"olpcfriends"?
Jan 21 12:46:03 <greebo> mchua, heh, the "one $device per $child
with $platform release team"
Jan 21 12:46:09 <cjb> I like it because it makes clear to me that this
is friends of olpc who are not actually olpc
Jan 21 12:46:20 <cjb> greebo: can we call it $$$ for short?
Jan 21 12:46:21 <mchua> Friends Of OLPC
Jan 21 12:46:24 <mchua> == FOO
Jan 21 12:46:31 <mchua> ...just sayi'n.
Jan 21 12:46:31 <cjb> and then we can have Foo Camps :)
Jan 21 12:46:34 <mchua> w00t
Jan 21 12:46:35 <m_stone> oy.
Jan 21 12:46:37 * Company has quit (Read error: 60 (Operation
timed out))
Jan 21 12:46:49 <m_stone> let's worry about naming later....
Jan 21 12:47:05 <m_stone> and just call it the
greebo+m_stone+... faction.
Jan 21 12:47:17 <greebo> cjb, I think olpc friends is ok, because
a) it sounds seperate to OLPC so we can encourage the OLPC refugees (and
annoyed peoples), it is low case olpc so hopefully avoids trademarks,
includes "friends" so hopefully sends a better message of inclusiveness
which everyone is looking for right now
Jan 21 12:47:23 <cjb> I think olpcfriends is probably a fine preliminary
name
Jan 21 12:47:39 <cjb> greebo: yup! those are all the reasons I like it
Jan 21 12:47:43 <m_stone> greebo: sounds fine to me as well.
Jan 21 12:47:45 <greebo> cjb, maybe not $$$ ;) people might think
we are a little too financially driven ;)
Jan 21 12:47:59 <cjb> if you're worried, shoot a mail to robert@laptop
asking if it sounds like a reasonable name for a grassroots deployment
group/company
Jan 21 12:48:01 <mchua> the longer I'm in this conversation, the more I
think I actually agree with everyone here, but am just using different
words and phrasings.
Jan 21 12:48:16 <cjb> I think that's exactly right
Jan 21 12:48:16 <m_stone> mchua: nah, I think there are still some
big disagreements lurking. :)
Jan 21 12:48:27 <greebo> "there is no cabal" <- re faction comment
from mchua :)
Jan 21 12:48:33 <mchua> m_stone: on methodology, but not on final
destination, I believe.
Jan 21 12:48:33 <cjb> my objections are all to things like "the phrase
'olpc release team' creeps me out" and not to anything you guys actually
want to do
Jan 21 12:48:48 * xobot has quit (Read error: 104 (Connection
reset by peer))
Jan 21 12:48:57 <mchua> cjb: yeah, I think most my objections are/were
to phrasings rather than actions as well
Jan 21 12:49:04 <mchua> phrasings and ways of stating things
Jan 21 12:49:09 <greebo> cjb, I've actually spoken to robert about
olpcfriends before, because it was originally going to be the grassroots
community devel just for the oceania region, will bounce this off him too
Jan 21 12:49:27 <greebo> m_stone, it's no fun if there aren't
disagreements ;)
Jan 21 12:49:28 <cjb> oh, this might help make my position even
clearer -- if olpcfriends decides that the Joint OLPC Release Wot Is
Made By Community People should be running debxo, I think that would
be a reasonable decision for it to take (as long as it was reached
democratically)
Jan 21 12:49:42 <cjb> since it sounds like people were worried I wanted
to force them to run/contribute to Fedora
Jan 21 12:49:49 <cjb> greebo: awesome
Jan 21 12:49:56 <greebo> cjb, gotcha, awesome
Jan 21 12:50:14 <mchua> cjb: the interface between
OLPC/they-who-make-ze-laptops + olpcfriends is an API to be worked out
as well
Jan 21 12:50:21 <greebo> k, cool, I'm going to get lunch now,
thanks everyone, this is a really awesome discussion, and I'll add more
documentation to the page later.
Jan 21 12:50:35 <cjb> yeah. thanks! we'd been needing to have a
conversation just like this.
Jan 21 12:50:43 * ctyler has quit (Read error: 110 (Connection
timed out))
Jan 21 12:50:43 <greebo> we already own olpcfriends.org so we
could start refactoring it to be global at some point :)
Jan 21 12:50:49 <mchua> (since having OLPC say "yeah, we'll ship
this as the factory image" is one big advantage of having the common
centralization)
Jan 21 12:50:57 <tinker-f595> hey mel
Jan 21 12:50:58 * ctyler (n=chris@global.proximity.on.ca) has
joined #olpc
Jan 21 12:51:02 <mchua> greebo.whereis() ? want lunch as well
Jan 21 12:51:12 <mchua> tinker-f595: howdy - was just about to pop
offline for lunch
Jan 21 12:51:16 mchua m_stone mako maquis Mercury morgs mpr mstrobert__
mtd mungewell1
Jan 21 12:51:29 <greebo> mchua, yeah, but that api is actually
quite simple. If OLPC want to participate in the process they have people
do that (same as HP have people participate in Debian). the only api
issue is the stupid keys issue, but general devel should be simple
Jan 21 12:51:31 <cjb> mchua: you're assuming there's going to be a
factory image :)
Jan 21 12:51:32 <mchua> cjb, m_stone, greebo: any objections to me
taking/cleaning/posting this transcript, for historical purposes btw?
Jan 21 12:51:47 <cjb> (there probably will, but I think the right
mindset is one that imagines that there isn't)
Jan 21 12:51:48 <greebo> mchua, I'm in social science building,
meet you at cafe?
Jan 21 12:51:54 <cjb> (or that it's 8.2.1 forever or something)
Jan 21 12:51:57 <mchua> cjb: oh, I thought ed wanted to have external
groups say stuff like "here, I want a million laptops with this image
on it"
Jan 21 12:52:00 <tinker-f595> mchua: have been talking to people on
anna's server about helping you test.
Jan 21 12:52:00 * Company (n=Company@169.222.10.65) has joined #olpc
Jan 21 12:52:06 <mchua> tinker-f595: *excellent*
Jan 21 12:52:09 <tinker-f595> that was a couple of days ago
Jan 21 12:52:19 <tinker-f595> at some point we need to know what
Jan 21 12:52:24 <cjb> mchua: that would probably work too, it just
gets complicated when the "million" part isn't there
Jan 21 12:52:40 <greebo> I think considering we can integrate
puppet into the laptops and schoolservers, the image that comes on the
laptop will be irrelevant soon enough (with 6 months perhaps)
Jan 21 12:52:43 <mchua> tinker-f595: I have to run off now - would
you mind shooting me an email reminding me to get to you on that? I'm
currently at a conference and the apt I'm staying in has no internet,
so my latency is kind of wonky atm
Jan 21 12:52:48 <cjb> mchua: transcript fine with me
Jan 21 12:52:52 <mchua> tinker-f595: thanks for starting up those
conversations, btw!
Jan 21 12:52:53 tinker-f595 Tales2 tarbo tglx ThomasWaldmann Thus0_ toi
Jan 21 12:52:55 <tinker-f595> we are remote form anna's server so no
mesh based collaboration it would be AP based collaboration
Jan 21 12:53:03 <mchua> tinker-f595: (and for your endless patience with
me on testing ;)
Jan 21 12:53:10 <greebo> so long as we can get OLPC to sign
stable images created by the community every so often, then we can easily
update/deploy using puppet.
Jan 21 12:53:11 <tinker-f595> but still collaboration
Jan 21 12:53:12 <mchua> tinker-f595: Ooh, that's actually fantastic.
Jan 21 12:53:20 <m_stone> greebo: soon even that won't be necessary.
Jan 21 12:53:29 <tinker-f595> well need to come up with a plan with you
Jan 21 12:53:29 <m_stone> greebo: see [[Partial key autonomy]]
Jan 21 12:53:41 <mchua> greebo: olpc is no longer going to be a bottleneck
for signing (that's the plan last I heard, anyway)
Jan 21 12:53:54 <mchua> s/is no longer/will someday soon not
Jan 21 12:54:00 <mchua> for some value of soon
