Working Group Minutes/EWG 2012-04-09

From OpenStreetMap Foundation

Attendees

IRC nick Real name
apmon Kai Krueger
tmcw Tom MacWright
TomH Tom Hughes
zere Matt Amos

Summary

  • zere asked for thoughts on a new time slot, but the channel was underpopulated (possibly because it's a bad time slot?)
    • zere to distribute a Doodle poll for a better time.
  • Google Summer of Code:
    • There is a proposal "Implement JSON REST API in OSM Rails port" [1], but it seems to be being dropped due to lack of support.
    • Problem with existing devs being mentors appears to be lack of time and/or uncertainty about time commitments needed.
    • There was discussion of whether spending time mentoring GSoC students is reciprocated by greater dev effort, or if GSoC students don't stick around after GSoC is over.
  • zere gave a license change bot status update.
    • It's not yet in a state where it can be run, even on an extract of the data.
    • Many unit tests still fail.
    • People can help by hacking the code, contributing unit tests or reviewing the test cases (no coding required) at [2].

IRC Log

19:01 < zere> apologies - i'm under-prepared for today... it really doesn't feel like a monday here ;-)
19:01 < apmon> too busy chasing bunnies?
19:02 < zere> painting the new flat. it's much more work than i thought it would be!
19:03 < zere> from last week - we were going to take a straw poll on more convenient times. i have nothing else for the agenda, so it's an open floor after that.
19:03 -!- Guest1314 [~ajturner@209.155.228.129] has quit [Ping timeout: 480 seconds]
19:03 < zere> so, straw poll: can you do an hour earlier (-1), current time (0), an hour later (+1) - if you can do multiple, then put multiple down.
19:04 < zere> e.g: for me, -1, 0, +1 - i could do any of those three
19:04 < zere> (on second thoughts, shall i just do this as a doodle poll and send it to the mailing list?)
19:06 < apmon> for me: -1 preferred (0 partly possible)
19:06 < apmon> Can I put GSoC on the the agenda?
19:08 < zere> sure
19:10 < apmon> Well, one question is can or should EWG be involved or help in any way with GSoC?
19:12 < zere> it seems like the whole change-of-time thing isn't getting a whole lot of attention, so let's move on to GSoC
19:13 < zere> if anyone has any change-of-time responses, just stick 'em inline and i'll do a count before the next meeting (or maybe a doodle...)
19:13 < zere> anyway, GSoC - what's needed? what's not happening?
19:14 < apmon> There is a GSoC proposal for "Implement JSON REST API in OSM Rails port". The question here is, is this something useful to the API?
19:14 < zere> yeah, that guy was posting on the dev list, right?
19:16 < apmon> I might have missed it, but didn't see it getting posted on dev
19:17 < zere> i thought there was a list thread somewhere about it
19:18 < zere> http://lists.openstreetmap.org/pipermail/dev/2012-March/024679.html
19:20 < apmon> Thanks for the link.
19:21 < apmon> At the moment, I suspect it won't be a priority to accept due to a lack of interest in the project from possible mentors
19:22 < apmon> zere: so if you think it would have any merit, it would be good if you could comment on it and see if the student would be interested in possibly contributing to cgimap instead
19:23 < zere> well, not instead - cgimap is an acceleration layer for the API, not a replacement for it.
19:23 < zere> so if we want to be implementing something in the API, it should go into rails as well.
19:23 < zere> we can totally prototype it in cgimap, but "the" API is the rails_port API.
19:23 < apmon> Well, then this could possibly be the opportunity for someone to work on the rails_port side.
19:24 < apmon> but it would need buy-in from the core rails_port dev's to be of any worth
19:25 < zere> i think it would be important to get a pretty serious proposal, and a mentor who knows about the API
19:25 < zere> s/the API/the rails_port code/
19:26 < apmon> Which is why I am bringing it up here. As if no one here is interested in mentoring it, chances to find an appropriate mentor would be slim...
19:27 < apmon> or at least a commitment to support the student
19:29 < zere> yeah. trouble is, i have no idea what the time commitment would be... if we're talking about a few emails / IRC chats here or there, that's pretty do-able. but i couldn't, in good faith, sign up to anything significant. i wouldn't be able to do it justice.
19:30 < zere> and i'm guessing from the general aura of silence that this meeting isn't necessarily the right place to be asking ;-)
19:36 < tmcw> Erm
19:37 < tmcw> As part of the editor work, if that works out, a JSON-based API will pretty much be a requirement
19:37 < tmcw> But GSoC has proven in the past not to be a very efficient way of doing things.
19:38 < zere> indeed. it's entirely possible that the amount of time spent mentoring would be greater than the amount of time to just implement it.
19:38 < apmon> my guess would be that partly the lack of interest from possible mentors does not help...
19:39 < zere> and partly that the number of people who really know the API & rails & all the necessary prerequisites is just too small...
19:40 < apmon> If no one puts in effort into training new people, then the knowledge will eventually just die out...
19:41 < TomH> most of the applicants seem to be applying for essentially mercenary reasons though rather than any great interest in osm 
19:41 < TomH> so I have no great desire to spend time mentoring them because the chances of them staying after GSOC are minimal
19:42 < zere> it depends on the person, but i guess the fact that they've not been seen around the community before doesn't exactly inspire confidence that they'll stick around after.
19:45 < apmon> Is there anything to discuss this week with regard to the license change?
19:45 < zere> maybe what might help is if someone who didn't know the API code, but wanted to, could mentor the GSoC student and, in turn, could be mentored by someone who knows the API code?
19:46 < zere> re: license change - it's just broken record stuff from me: if anyone would like to pitch in on the license change code on github, it would be greatly appreciated.
19:48 < apmon> The last osmf blog post said that there was some testing ongoing. Are there specific things that need tackling from those tests?
19:51 < zere> frederik pushed some tests based on real-world examples "in the wild"... and they failed.
19:52 < zere> so i could use some help tracking down why they're failing - whether a bunch of them have a common cause, that sort of thing
19:52 < zere> and, of course, any review anyone wants to do - pointing out weirdnesses, inconsistencies, etc...
19:53 < apmon> Is it in a state that can be run on an extract of the data (even if it doesn't always give the correct results)?
19:53 < apmon> Then people could look through the results and check they are OK and report all of the cases that aren't
19:54 < apmon> This would allow non coders to help
19:57 < zere> non coders can already help - the test cases should be pretty close to human-readable. if not, we can easily do some more documentation.
19:57 < zere> and no, the code isn't in a state where i could run it on an extract yet.
19:57 < zere> so help with that would be good too :-)
19:58 < apmon> :-)
20:02 -!- ajturner is now known as Guest1319
20:02 -!- ajturner [~ajturner@209.155.228.129] has joined #osm-ewg
20:06 < zere> ok. this seems to have run its course - hope to see you all next week, same time, and i'll make sure it feels like a monday ;-)