Oversight Board/2010/Meeting Log-2010-08-24

From Sugar Labs
Jump to: navigation, search
<walterbender> oops. sorry I am late
anyone around for the SLOBS meeting?
  • cjb: is
<walterbender> Adam and Bernie both said they'd be here :)
cjb: how was LinuxCon? (I missed most of it :()
<cjb> good! I was surprised, it's generally been non-technical, but not this year
<walterbender> cjb: I wanted to go tot eh git sessions
cjb: but I was busy patching Turtle Art :)
<cjb> they were pretty intense. I can point you to the slides, though.
<walterbender> cjb: great.
<cjb> oh, yeah, he used a weird presentation program. http://github.com/schacon/showoff-wrangling-git is the repository for the slides, but there's some ruby code involved in presenting them.
<walterbender> cjb: well, he could have used Turtle Art :)
<tomeu> hello
<walterbender> hi tomeu
tomeu: still no quorum :(
while we are waiting, I had been thinking for a while about getting a small group together to try to write up a Sugar document equivalent t the Bitfrost document... something that lays out a longer-term vision of our goals, not the means by which we achieve them. do you think this would be a worthwhile effort?
<tomeu> well, my position has been that we need to grow capacity before we can do such stuff properly
not that I have had much success convincing people to discuss about resourcing
=-=: unmadindu is now known as unmad|away
<walterbender> I think that a manifesto might attract more resources to the project, if done well
<tomeu> as we stand today, I see it quite likely that we won't get enough people with experience and responsibility to find time to write that
so we are likely to get wishful thinking
<walterbender> tomeu: but, yes, it will tax our most experienced people...
tomeu: a bit of wishful thinking is OK... Bitfrost is far from being complete in its implementation
<tomeu> walterbender: I see interesting that you relate it to the resourcing question. if it's a proposed solution, maybe we should first try to frame the problem and comapre to other propable courses of action
<walterbender> but the existence of the document helps more quickly focus discussions/debates
tomeu: so I see it saving resources in some respects as well
<tomeu> walterbender: maybe you want to explain your idea in more detail and specifically how it relates to the resourcing question?
<walterbender> I agree... given limited resources, we should look at other proposals as well.
tomeu: well, we spend a lot of time debating both what to do and how to do it.
tomeu: and the extent to which the debates can be made more efficient, we can save time and energy
tomeu: and if we have a clearly defined taxonomy of goals, new people can jump in and explore ideas in a way that is likely to be more congruent with our needs
more likely their proposals will be better grounded and their code useful to us.
<tomeu> to be honest, this doesn't matches what I have observed in our community
<walterbender> this is speculation of course.
(on my part)
<tomeu> do we agree we have a problem with the lack of maintainers?
<walterbender> tomeu: yes...
<tomeu> and do we agree we don't have a plan to solve that?
<walterbender> tomeu: maybe the first step is to reiterate where the resource problems lie
<tomeu> I think we have never discussed that frankly enough
<walterbender> tomeu: then let's go for it
I agree it is a more pressing issue. but I am not sure what action to take
<tomeu> walterbender: we know some members of our community have resources to contribute
for months they have stated this will, but this hasn't ended up in new maintainers
<walterbender> tomeu: I think that they still don't understand the details of how to help
<tomeu> we have been waiting for them to step up and maybe that's not the best strategy
walterbender: I agree
so SLs should provide a way forward
so the problem could be framed as how to help those that want to take responsibilities inside SLs
this can be split in: * specify which responsibility is waiting for a taker, * which experience needs every vacancy, * how can people acquire that experience
AC is working on this same problem on their own
<walterbender> tomeu: I think the degree to which we can be explicit in our request, .uy, .pe, etc. will allocate resources...
<tomeu> we can leave them go their way, or we can join, or whatever
walterbender: I also believe that some will, but planning this in a credible way will take some resources
and coordination
<walterbender> tomeu: Catch 22
<tomeu> I have tried to do it alone, but cannot focus on it, even less if others don't agree this is important and are willing to pool their time
-->|: CanoeBerry (CanoeBerry@d5rmxxf1.media.mit.edu) has joined #sugar-meeting
<tomeu> so if we are still trying to do non-trivial things in SLs, we should really put what it takes and do it
other FOSS communities can do things that go past individual efforts, we should be able to
the alternative to that is to just wait and see how OLPC and AC do that job
they don't seem to be doing bad at all
they are hiring people and putting them to be maintainers
I would recommend to wait until we have a more balanced community in terms of maintainers and code to maintain before we decide to write such a document
<walterbender> tomeu: OK. thanks for the feedback.
tomeu: I think we cannot be wholly dependent on AC and OLPC... we need deployments to jump in too.
<tomeu> the problem is similar, we need a more coordinate action so we can make a credible proposal to deployments
so they displace a bit of their work into our community
again, is something that requires more than a single individual's spare time
you know I have tried this as well several times, without success
=-=: bernie_afk is now known as bernie
  • bernie: arrives
<cjb> bernie: dude
<tomeu> but bernie is solving all those problems single-handedly, so we can just let him do his job
<walterbender> tomeu: :)
bernie: do you have the backlog available?
<bernie> tomeu: lol :-)
too bad I'll be back in boston in october...
walterbender: yes, just finished reading
walterbender: sorry, I was visiting a school outside montevideo
<cjb> bernie: it's okay, you're only 58/60 minutes late.. :/
<bernie> cjb: better than last time
I was very impressed: teachers are very engaged and not afraid of technology
kids ask very smart questions. they've seen the XO-HS and asked if it fixes the touchpad :-)
they also said they wanted more memory and a faster processor than the XO-1. We had good news for them
ok, enough spamming the meeting
please go ahead
<walterbender> bernie: if you get a chance to meet Rosale, a teacher from Duranzo... she is great!!
bernie: tomeu and I were talking about how to grow the ranks of our maintainers...
bernie: your experience would be useful input
I should also start logging...
#startmeeting
<meeting> Meeting started at 12:04 UTC. The chair is walterbender.
Commands Available: #TOPIC, #IDEA, #ACTION, #AGREED, #LINK
<bernie> walterbender: I was in Salinas. They use Scratch, Etoys and some Turtle Art. They say they need more activities for teaching math and that geography now is much more engaging thanks to wikipedia / google maps
<walterbender> a quick here to record our quorum?
  • walterbender: here
<tomeu> here
  • bernie: here
  • cjb: here
<bernie> mchua_afk, CanoeBerry: ping?
<cjb> don't know how long I can stay, not at 1cc yet
<bernie> anyone knows what happened to sean?
<walterbender> CanoeBerry: you arrived!!
<bernie> he's been very silent lately
<walterbender> cjb: as long as you can :)
#topic maintainers
<tomeu> bernie: maybe you want to make a summary of your efforts in bringing new maintainers?
-->|: tch (~tch@mail.paraguayeduca.org) has joined #sugar-meeting
<walterbender> bernie: any comment on the backlog discussion (which I will add to the log)
<bernie> tomeu: yup
so, I think paraguay got to the point where it could autonomously develop sugar. it took some time with dsd and then some time with me
however, something went wrong
as you may know, rgs left one month ago
then me
<cjb> oh, I didn't know about rgs
<bernie> the remaining people are overworked and stressed due to lack of a barrier between them and the field. they might quit as well
so I'm very sad to say that we cannot count on Paraguay Educa doing too much engineering from now on
I'm now in uruguay. I have seen huge progress from no participation at all to a good level of engagement. technical skills have also gone up a lot
they're hiring new sugar developer(s)
<walterbender> I guess it just takes time...
<bernie> nepal has lost its CTO (bryan) and has hired someone (peter something) to continue development of Karma
<walterbender> but we should also try to understand what is broken in .py... a big investment in engineering and now it is vanished?
did we do something wrong?
<bernie> there was some interest on your side to update their platform (dextrose) but I don't think they plan to put engineers on Sugar itself
<walterbender> or was it circumstance?
<bernie> walterbender: I think it was just unlucky circumstances. both me and rgs leaving at the same time have disoriented the others.
walterbender: as you know, there was also a strong conflict between a very innovative techteam and a very conservative eduteam
<tomeu> bernie: and on the AC front? how are they progressing on the maintenance front?
<bernie> walterbender: in uruguay the "capacitadores" seem much much much more technical and would not have lobbied to stop a software upgrade
tomeu: AC has been struggling to find skilled full time engineers
tomeu: m_anish is a star.
<tomeu> and aside m_anish?
<bernie> tomeu: aa and alsroot are also participating, but the first has very little free time and the second just started
<walterbender> bernie: Ironically, from the pedogogical POV< the .py edu team is about the most progressive I have seen
<bernie> tomeu: david also temporarily hired rgs to work on Browse + WebKit, but after a couple of weeks he decided he'd rather stay away from the entire olpc world for a while longer.
walterbender: yes, that's true. pacita is amazing from the pedagogical pov
<tomeu> bernie: so from your side, there isn't really interest presently from deployments in providing maintainers?
<bernie> walterbender: however, until now, she has never used sugar herself. (I've not seen her since he went to boston, maybe you changed that?)
<walterbender> tomeu: interest or ability?
<tomeu> walterbender: interest
<walterbender> tomeu: I think there would be interest
<bernie> tomeu: I think they don't know what it means exactly and why they'd want to maintain something when they could simply change random things
<tomeu> walterbender: I have heard in the past about interest, but I'm not sure it's still valid
bernie: ok, that's a very interesting point
<walterbender> tomeu: I think it is (1) not having a clear understanding and (2) a lack of experienced people
<bernie> tomeu: pyedu, with its free-software-aware cto, was the most likely to grow a maintainer quickly, but now I wouldn't hold breath any more.
<tomeu> I was hoping that through their involvement in the non-technical side of things in the deployment team, they would get to know progressively how their software is made
walterbender: yes, I'm tying to understand a bit better in detail
<bernie> tomeu: I think david might be interested to put someone full time at maintaining things. for example, m_anish.
<tomeu> bernie: that would be great
<bernie> tomeu: david's short term goal is to send some of his men to deployments so they can build expertise on the needs of deployments.
<tomeu> bernie: so the biggest blocker is making deployments aware of how their software is made and how it affects them?
<bernie> tomeu: this is the part of his plan that I like the most.
<tomeu> bernie: and deployments learn from them?
<bernie> tomeu: that's hard: they are at a stage in which their resources are split between: logistics, hw support, network infrastructure, schoolserver, platform work, sugar and activities
tomeu: if they do employ 5 decelopers, it's unlikely that one would be put full time hacking on sugar.
tomeu: activities are perceived (note, perceived) as more important to pedagogy than sugar itself.
<tomeu> well, ideally none would be full-time
<bernie> tomeu: the view that the laptop is just an array of activities is, of course, naive... but it might take one more year before they see this.
<tomeu> bernie: hmm, what will happen during that year?
<bernie> tomeu: it takes more expertise to understand the role of the environment than the hardware or the individual activities...
tomeu: it's more subtle (and more fundamental too... but how do you explain it?)
tomeu: my hope was that uruguay and paraguay would continue their technical collaboration and maybe add more developers
tomeu: working on the platform is a perfect excuse to do plenty of sugar work as well, as we did in dextrose
tomeu: however, the quality of their contribution is still somewhat below the level we expect upstream.
tomeu: there's awareness of the importance to upstream changes, but it's not sufficient to overcome the barriers.
<tomeu> bernie: do you think we cannot expect to find a way to raise the quality of those contributions in the med term?
<bernie> tomeu: some patches were written with variables in spanish and there's close to 0 gain for the deployment to rewrite them after they have already gone in production.
tomeu: I think we need to create an ascending slope like we do for children... just raise the pole gently as people get more experienced...
<tomeu> bernie: that's interesting, so we could try to create a guided progression?
<bernie> tomeu: because I had only junior-level contributors (except for tch) to work with me, I had to choose between not doing anything or allow dozens of unupstreamable patches in dextrose
tomeu: the risk, of course, was to fork too much, which is what happened. we have some 100 unmerged patches, of which maybe a dozen are going to be merged soon.
well, maybe two dozens
tomeu: now that the release pressure is down, tch, aa and m_anish are spending some time upstreaming their work
<tomeu> guess the dextrose schedule didn't played that well with 0.90s
<bernie> some patches really aren't worth upstreaming at all... they are deployment customizations such as the reintroduction of the Restart button
tomeu: oh, yes. that was really unfortunate...
tomeu: again, there was a hard choice to make: either align with the school year and do a release by august, or align with upstream for better collaboration
<tomeu> sure, that's what I guessed
<bernie> We should probably get used that the former will always take precedence in all deployments...
I'm not suggesting we adjust the sugar release schedule to follow through... we might want to be prepared to have always a big delta between upsteam and what gets deployed
as long as patches flow upstream at the same speed they are developed, there's no problem
<tomeu> sure, downstreams will get better at managing their delta
well, at least looks like who are putting more development resources are also providing maintenance help and also guiding themselves by what deployments need
<bernie> anyway, there are two types of fork. the hostile ones and those meant to accommodate different needs of different user bases... the latter is what we do and it's not necessarily a bad thing
<tomeu> so we are in a sustainable path even without deployment's direct intervention
so maybe SLs can do without them for a bit longer?
<bernie> tomeu: indeed. collabora and olpc are putting some good engineers on sugar
plus AC...
the goal of Sugar Labs (in my mind) was to create an industry around Sugar... and this seems to be happening finally
<tomeu> yeah, m_anish is good and alsroot is going to a deployment
bernie: I still think it would be more efficient for deployments pooling resources in SLs, but maybe that will come later
<bernie> the disconnection between deployments and upstream development will be filled if 2 more developers besides me get to travel around
<walterbender> bernie: I've never been successful raising money for that :(
<tomeu> walterbender: well, from what I have heard deployments would be happy to pay the costs
the problem is rather finding talented people that can do that and also not get paid as much as they could
what SLs should do is to organize (and fund) a sugarcamp
that's what GNOME Found. mainly does
<walterbender> tomeu: we have an open invitation to hold a Sugar camp in Miami...
<bernie> walterbender: perhaps funding within SL is not what we actually needed to keep development going.
walterbender: unlike many, I strongly believe that even a 100% volunteer-driven project can go a long way. what was missing in SL so far was contact with users.
<walterbender> bernie: not funding developers... funding face-to-face time
bernie: so I think we are in agreement
<bernie> walterbender: oh yes. that's really important
<tomeu> walterbender: ah, I was seeing as two different things having developers spending some months in a deployment, and gathering people in hackfests
what bernie has done is more the former
<bernie> tomeu: this might sound shocking at first, but if you think about it, *today* almost all the sugar contributors are paid in one form or another.
tomeu: notable exceptions were sascha and aleksey, but we're fixing the latter at least.
<walterbender> bernie: paid to work on Sugar or paid?
<tomeu> bernie: old news :p http://blog.tomeuvizoso.net/2010/01/sugar-labs-has-changed.html
<walterbender> bernie: and paid a competitive rate?
<bernie> walterbender: well, I've been paid (although very little) along with tch and jasg. tomeu and erikos are both paid... all the AA folks too...
walterbender: plus olpc is hiring a bunch of full-time and part-time positions...
walterbender: we seem to have lost one paid developer, though: sayamindu. which will be VERY missed.
tomeu: yep, I remember reading that blog entry of yours. I liked it very much.
<cjb> bernie: this is pretty good, it means we can say "hack on sugar, there's likely to be a paid job in it for you if you're good", etc
<bernie> so the problem of not having paid positions is mostly solved. actually, deployments and AC can't seem to find more sugar developers easily.
<tomeu> yeah, I'm worried about finding good people to cover positions
<bernie> cjb: I was almost sure that would have happened if the install base of sugar became large enough
<walterbender> tomeu: I agree
the people bernie listed are all already participating...
<bernie> cjb: 2 years ago, the olpc installation base was way too small and the early deployments were still fighting hw and infrastructure issues... now they are mature
<walterbender> we've seen little increase in our numbers through $$
but maybe avoided some attrition
<tomeu> bad news are that companies are paying an awful lot of money for people with the required skills
<bernie> walterbender: this is normal, the linux distributors *always* hires from the community.
walterbender: this made redhat and canonical leaders, while sun and novell loosers (because their community relationships were terrible)
<walterbender> good thing we are so adept at community relations :)
<tomeu> heh
so that's the third issue
didn't had any luck with the community team or manager
<bernie> our community seems healthy to me... not so much SL as an organization, though. are you guys as worried as I am?
we seem to have lost also sean now...
<tomeu> bernie: I'm not as worried with SLs because the sugar ecosystem seems to be doing well without us :p
<bernie> at the organizational level, very little is happening within SL lately.
<tomeu> but it's a bit sad to think of the lost opportunities
<walterbender> tomeu: such as...
<bernie> yes, I wouldn't be particularly bothered if it turned out that SL has reached its goal and is no longer important to the future of Sugar.
<tomeu> walterbender: I said above that things are moving in the right direction, but because of what are doing people not directly involved in SLs
such as AC, OLPC and Collabora
SLs has had a passive role in the success of those organizations
<bernie> tomeu: I liked bemasc's comment a few weeks ago that we should hand off SL to deployments.
of course, it cannot happen by signing a contract.
it would have to happen gradually, ideally by making deployments people escalate SL from the inside...
however, I don't see much interest from deployments in covering the "political" roles of sugarlabs: they don't candidate for slobs, team coordinators.. not even mere members.
<tomeu> walterbender: to be frank, I haven't felt supported enough when I tried to start the deployment team, nor the community team, nor starting the maintainer training program
<walterbender> tomeu: Maybe in the case of Collabora there is no direct involvement in SL, but in the other cases, it is hard to tell where one begins and the other ends...
tomeu: I am not sure what we can do to provide more support in those cases...
<tomeu> walterbender: coordination
<bernie> tomeu: at that time, I wasn't a deployment person. if you did it now, I would have participated.
<cjb> walterbender: oh, could you talk about the SLOBs election? when's it supposed to happen?
<tomeu> bernie: maybe we just need to try from time to time
<walterbender> cjb: good point...
<bernie> tomeu: rgs tried to revive the Deployment Team, but after one interesting meeting it went dark again.
<walterbender> can we switch topics?
<bernie> sure
<walterbender> bernie: I think rgs__ was too busy with other things... we should not have agreed to let him take it on...
<bernie> tomeu: (for later) have you seen what marcopg has been up to? sugar-core, I mean.
walterbender: more than agreed, we kind of pushed him to do it
<tomeu> bernie: bits of it
(not suire if I have missed anything)
<cjb> looks like last year it ran 2-16 Oct or so
<walterbender> #topic election
cjb: what I don't recall is if we adopted the staggered term concept or abandoned it
cjb: I think the election mechanics worked out OK last year... better than the first time
<bernie> walterbender: indeed
<walterbender> cjb: if we use Sept. to solicit candidates, we can have an election in Oct. again
<bernie> I'm not sure whether I would candidate or not... I'm afraid that next year my role within SL will diminish too much to justify being a slob.
<walterbender> bernie: :(
<bernie> otoh, now I have more deployment experience, therefore I could represent deployments within the board.
<walterbender> +1
<bernie> ideally, I'd like a real deplyment person such as rgs to be on the board instead of me.
<cjb> walterbender: should we ask for a volunteer to run the election, or maybe get a SLOB who doesn't intend to stand again to do it?
<walterbender> be nice to recruit someone from the pedagogy side
cjb: either is OK... maybe the latter...
cjb: I can ask lfarone if he'd be willing to do the membership drive again
<cjb> sounds good
<bernie> walterbender: this is something I always think about: how to attract more pedagogists within SL.
walterbender: lfaraone did a great job last time
<walterbender> bernie: they are out there... the weekly chat is getting good attendance
<bernie> walterbender: who was the other member of the membership committee? quozl?
<walterbender> bernie: I don't recall...
<bernie> walterbender: weekly chat of claudia urrea?
<walterbender> si
we have a few action items to take care for the election:
1. get the membership committee in order
2. recruit someone to run the election
3. solicit some candidates
4. hold the election
<cjb> walterbender: it looks like (at least from the email record) we did do staggered, top four for two years
<walterbender> cjb: great... makes it easier... but who were the top four?
<cjb> http://www.olpcnews.com/use_cases/community/sugar_labs_oversight_board_election_results.html
walter, tomeu, mel, bernie
<bernie> damn, I don't need to recandidate then :-)
<walterbender> so bernie is with us next year whether he likes it or not :)
<tomeu> bernie: change of plans! :)
<bernie> lol, it's my destiny :-)
whether or not I'll be a slob , I can *guarantee* continued support of SL from within the infrastructure team.
with SMParrish and rgs__ just joined in, the SL infrastructure is really well staffed this year.
<tomeu> bernie: infrastructure is really fundamental
\o/
<walterbender> any ideas about #2 above, or should I just solicit in the Digest?
<cjb> or an iaep@ mail. I expect someone'll volunteer..
<walterbender> I am asking lfarone re #1 on #sugar
<cjb> cool
I should go to work now, will catch up later
thanks
<walterbender> cjb: thanks. CU
<bernie> cjb: you can't run the election this time because you'll have to be re-elected.
cjb: too bad, I liked how you did run the election last time
cjb: ah, you've dropped off the meeting
<cjb> bernie: I didn't do it last time :)
I think dfarning did
<bernie> cjb: really? I was sure it was you.
<walterbender> I guess since I am not standing for reelection, I could do it.
<bernie> walterbender: +1
<walterbender> seem OK?
but I will try to solicit someone nonetheless...
do we need any motions re this discussion? seems administrative, not legislative.
Anyway, we should wrap up the formal meeting...
unless there are any other pressing issues?
If we meet again in two week? I hope to have the election plans sorted out in the interim.
OK
3
2
1
#endmeeting
<meeting> Meeting finished at 13:22.
Logs available at http://me.etin.gs/sugar-meeting/sugar-meeting.log.20100824_1204.html