22:59:10 <mchua> #startmeeting
22:59:11 <zodbot> Meeting started Tue Feb 23 22:59:10 2010 UTC.  The chair is mchua. Information about MeetBot at http://wiki.debian.org/MeetBot.
22:59:13 <zodbot> Useful Commands: #action #agreed #halp #info #idea #link #topic.
22:59:30 <mchua> All right - this is for me to take down my responses while reading the log from quaid.
22:59:34 <mchua> #link http://meetbot.fedoraproject.org/fedora-gsoc/2010-02-22/fedora-gsoc.2010-02-22-17.04.log.html
23:00:15 <mchua> I'll start by noting that I was hanging out in #gsoc last night, since I just got back from RIT and Steven Jacobs (a professor there) had some questions about RIT students participating in GSoC that seemed good to answer in a public channel.
23:01:00 <mchua> Leslie Hawthorn was (as usual) there, and we talked briefly about making the GSoC model independent of the G - basically, what we're trying to do here - and she was a strong +1 to it and wants to be kept in the loop.
23:01:17 <mchua> So we're generalizing it with her blessing, so to speak.
23:01:38 <mchua> Now, reading through the log... Fedora has a long history of success with the Summer Coding model.
23:02:28 <mchua> and GSoC in particular. And seems to have done a good job of capturing that knowledge - which we get to learn from now for the Next Step Forward, because it also sounds like we've plateaued.
23:03:48 <mchua> It sounds like we want to do two things for this Next Step - the first one is the aforementioned "how do we make this model independent of the G in GSoC"?
23:04:01 <mchua> so making that dependency an optional, swappable part.
23:04:52 <mchua> And the second one being "how do we increase our yield of quality student successes at the end of a particular Summer of Code"?
23:05:52 <mchua> Which could be done by increasing the number of attempts, the percentage of attemps that end in success, etc. - bunch of metrics - but the end goal is "more successful students projects at the end of summer, for a less-than-linear relationship to investment input
23:06:15 <mchua> quaid: that sound about right?
23:06:33 <mchua> so this week, while we're swapping off, is kinda like Marketing Week for our SoC stuff
23:06:50 <mchua> the meat is there, thanks to the FAD; we just need to do the plating and advertise that supper's going to be ready.
23:07:03 <mchua> quaid: a couple questions for you.
23:07:10 <mchua> * are you running GSoC for Fedora (and JBoss) this year?
23:07:26 <mchua> * if so, are you planning on running it again next year - or is this year a handoff year?
23:07:42 <mchua> * are there any non-G SoC programs going in Fedora this year (or should there be?)
23:08:28 <mchua> * aside from the FAD, what has the summer coding sig been doing so far? (Though I'll read the list archives/logs/wiki as well, when I get back from dinner).
23:08:44 <mchua> I think I have enough to start running slowly, though.
23:08:51 * mchua --> food, bbiab
00:43:13 <rrix> mchua: yargh, connection problems?
07:54:41 <quaid> for mchua's log ...
07:55:31 <quaid> I think the make it a model independent of Google is important, but cannot get in the way of a succesful G-as-in-Google summer effort
07:55:41 <quaid> mainly because I'm a'feared of splitting sarce resources
07:56:02 <quaid> so I'll personally put off for a year the generalized approach, if it threatens the 2010 GSoC
07:56:06 <quaid> that said ...
07:56:14 <quaid> it sounds as if there are people who want to help with that in particular
07:56:39 <quaid> so we can do it, with the caveat that IMO Fedora and JBoss resources should focus on the 2010 experience mainly.
07:56:46 <quaid> now, to answer questions ...
07:58:30 <quaid> 1. I have been *an* admin for GSoC for years, and last few years the main thing I did was marketing within Fedora, but I had a few saves; it's worth having a few hands paying attention, it's a lot of details.  I'd like to continue being an admin. Patrick Barnes (nman64,pwbarnes) has been the GSoC admin as far as I recall, and has focused a lot on the meat-and-bones of the process
07:58:42 <quaid> also, Bob McWhirter was admin for JBoss.org for the last few years
07:59:12 <quaid> sankarshan_away has _always_ been King of Mentors esp. with reviewing applications, and I'd like to see him being an admin this year, if he cares to
07:59:23 <quaid> one reason is, admins have an explicit role we didn't have in the past
07:59:41 <quaid> in the past, admins for our efforts didn't exert buck-stops-here authority
08:00:09 <quaid> but in the model we learned from KDE et al, the admins resolve conflicts with decisions when they have to be made, and often apply their own metrics
08:00:31 <quaid> the admins the the tool-wise power to resort the students; sometimes the admins know or believe something that makes a resort worth it, etc.
08:00:49 <quaid> there is a tradition here and folks say it helps make their effort better
08:00:54 <quaid> we've written about it ...
08:01:02 <quaid> but we don't want people to be surprised about that
08:01:33 <quaid> so we want those "expanded" powers highlighted for mentors; not really expanded except in being more explicit.
08:02:53 <quaid> 2. I don't now of any general Summer Coding programs in Fedora this year, although ... what are Fedora interns but summer coders directly under our management?  So, include interns in your thinking.  (I've long argued for Engineering managers to consider how to get their teams to get GSoC student projects.)
08:03:05 <quaid> with a tie-in to other interns
08:03:38 <quaid> the *idea* of the [[Summer Coding SIG]] is to be that meta-framework for "stuff that is SoC but not G"
08:03:46 <quaid> and the way to get there is through an improved 2010 GSoC
08:04:16 <quaid> we can then say, "See how much more ass we kick now under this new model," and attract more attention for the generic
08:04:30 <quaid> generic should go all the way up to influencing or creating a greater mentoring program for Fedora
08:05:04 <quaid> but ... if decause sees a good intersect, we can devote time _from_our_team_ (initially) to pair-up and make that happen
08:05:11 <quaid> likely we
08:05:21 <quaid> we'll see other ideas once the framework is in place
08:05:40 * quaid ends answering question 3
08:05:43 * quaid ends answering questions
08:09:58 * mchua reads backlog (thank you, zodbot)
08:11:42 <mchua> quaid: so basically my deliverables for the week are https://fedoraproject.org/wiki/GSoC_2010 and [[GSoC 2010 sub-project how-to]] finished, with a bunch of posts to lists/blog about 'em with the SIG in the loop?
08:13:20 <quaid> mchua: that's basically it, I think
08:13:24 <quaid> and helping to make sense of the holes
08:13:32 <quaid> there are still parts I'm concerned aren't visible
08:13:44 <quaid> e.g. that long bit about expanded admin powuh and authoritah!
08:14:21 <quaid> last year the situation with JBoss.org was akin to "surprised *and* disappointed the JBoss.org mentors"
08:14:40 <quaid> so we don't want them to miss the way things are and the protections in place to keep from gaming the system
08:14:52 <quaid> for example ...
08:15:06 <quaid> instead of it being "this pool of Fedora v. this pool of JBoss"
08:15:29 <quaid> it's "all of these sub-projects competing for the limited resources"
08:15:48 <quaid> with the sub-projects simply within one or the other of Fedora/JBoss.org
08:16:36 <quaid> the point is, give us all something to really vote for and we'll earn student slots, not by gaming with bullshit projects to try milking for slots.
08:16:56 <quaid> (not that anyone did the latter, but that's the next spot down the slope when one starts gaming Melange ...)
08:17:47 * mchua nods
08:18:26 <mchua> Well, I'll start running on this tomorrow after POSSE stuff and will holler after a first pass - I've got at least enough to leap out of the starting block now.
08:19:09 <mchua> And with that, I'm going to close the logs. :)
08:19:12 <mchua> #endmeeting