Working Group Minutes/EWG 2012-10-08

From OpenStreetMap Foundation


IRC nick Real name
alexbarth Alex Barth
apmon Kai Krueger
pnorman Paul Norman
ppawel Paweł Paprota
RichardF Richard Fairhurst
shaunmcdonald Shaun McDonald
tmcw Tom MacWright
TomH Tom Hughes
zere Matt Amos


  • Top Ten Tasks review
    • the list of tasks is at [1]
    • ACTION: TomH to push current OSB/notes branch public (which he has been improving on a local branch)
    • (OSB/notes) there are concerns about the UI implementation of the current branch, but it may be possible to merge the API separately so that UI development can proceed independently
    • AGREED: PL2 tutorial mode doesn't seem like a priority, given the expected move away from Flash as a technology, so we'll move it out of the TTTs
    • ( i18n) [2] "About the multiple site, yes is possible, we currently have 5 sites running over the same install, including this one, but is still not possible to share a database, but it will be in a near future I promise"
    • ( i18n) ACTION: ppawel to look at feasibility of migrating OSQA to shapado (or some other OSQA alternative)


18:01:07 <zere> #startmeeting
18:01:07 <ewg-meetbot> Meeting started Mon Oct  8 18:01:07 2012 UTC.  The chair is zere. Information about MeetBot at
18:01:07 <ewg-meetbot> Useful Commands: #action #agreed #help #info #idea #link #topic.
18:01:40 <drol> #help
18:01:56 <zere> welcome to the re-boot of EWG. on the agenda today: review of the top ten tasks and events planning
18:02:16 <zere> so let's kick off quickly with the TTT review
18:02:36 <zere> #topic Top Ten Tasks review
18:02:48 <apmon> For reference:
18:03:17 <zere> #info the list of tasks is at as apmon says
18:04:23 <zere> item one on the list is the OSB/notes branch. has there been any action there since the last time this was reviewed (6 months ago?)
18:04:48 <TomH> I have done some work on it recently - hope to have it back out in public soon
18:05:15 <zere> awesome. what's the remaining work to be done?
18:05:40 <TomH> well I can create new notes from the web interface again now
18:05:52 <TomH> just need to be able to add updates to existing notes I think
18:06:22 <apmon> Would it be possible to commit what you have anyway, even if it is not finished yet?
18:07:03 <tmcw> Yeah, I think that'd be helpful - at least I've got very little idea of the code behind osb
18:07:33 <zere> i know that alexbarth and the rest of the mapbox folks were interested in whether the workflow could support the kind of things they're currently doing with github
18:08:19 <zere> i.e: bugs/notes could be claimed to work on, commented on, watched, etc...
18:09:24 <apmon> Imho, we should get a basic version working and merged first. Then more elaborate work flows can be added afterwards.
18:09:26 <TomH> well frankly I'm not at all sure about the whole thing - the basic API is fine but the whole web interface bit is really a bit of a nightmare
18:09:48 <zere> in what way?
18:09:49 <tmcw> Well, it'd best be a public nightmare
18:10:11 <TomH> I'm starting to come to the conclusion that trying to do it all in OpenLayers popups is a mistake and we should open the side bar to show the details and do editing etc
18:10:27 <ppawel> isn't it public already? ->;a=shortlog;h=refs/heads/openstreetbugs
18:10:30 <zere> apmon: agreed, as long as the basic version isn't going to give us horrific upgrade problems
18:10:31 <RichardF> I suspect there are more JS/UI hackers within our community than Rails/db hackers
18:10:35 <tmcw> pp
18:10:43 <TomH> ppawel: that's way out of date
18:10:45 <tmcw> ppawel: I think that's behind TomH's version?
18:10:50 <ppawel> OK
18:11:03 <pnorman> btw, top ten tasks is currently top nine I think
18:11:19 <TomH> tmcw: I'm reluctant to push it out there and have people starting adding to it before the basics are complete - trying to stop it become a complete mish mash
18:11:40 <apmon> RichardF: Yes, as we have seen by the recent clean up commits to javascript from several people.
18:11:49 <zere> pnorman: nah, i resurrected the closed ones, so it's still 10. they're down the bottom under "closed tasks"
18:11:52 <tmcw> I get the concern about making it public, but don't think that this is sustainable.
18:12:24 <TomH> the whole thing is a cockup really - we started in completely the wrong way and have wound up basically rewriting everything without actually just giving up and restarting
18:12:25 <tmcw> Like, we need to have some clarity to know how much work needs to be done, what the architecture is looking like, what's up with the "Leaflet" version, etc.
18:12:46 <TomH> what leaflet version/
18:12:58 <tmcw> it's totally fine if osb has problems; but like there's this disparity behind the continual 'ready to merge!' message and the 'no, not open source yet' message
18:13:02 <TomH> it's all built on OL like the rest of the site
18:13:30 <apmon> Please don't add leaflet to the equation for now, or it won't ever be completed
18:13:35 <tmcw>, right?
18:13:45 <tmcw> apmon: not adding it to the equation, it
18:13:50 <tmcw> is a thing that already exists
18:13:50 <TomH> no idea waht is
18:13:54 <tmcw> Haha
18:14:00 <TomH> nothing to do with us AFAIK
18:14:13 <RichardF> YAAttemptToBuildAnOffsiteOSBService
18:14:21 <tmcw> and they have their own, mercurial repo!
18:14:25 * tmcw community!
18:14:50 <pnorman> I think uses the same bugs DB as the others
18:15:26 <zere> source code links to here: which suggests it's a different db at least
18:15:28 <tmcw> Like, right now the MapBox folks are like, what should we think about as far as issue trackers.
18:15:54 <apmon> TomH: What can we do to help move this along?
18:15:58 <ppawel> TomH, would it make sense that you write down all your major technical (and functional) concerns with this topic? sounds like there are some real showstoppers...
18:16:09 <tmcw> This mishmash is not a very good thing for that, at the moment, it's like working on an out-of-date public repo, or waiting for someone to merge, or starting on a 'brand new nih-syndrome' project
18:17:04 <ppawel> perhaps as an update of
18:17:56 <zere> perhaps the first question is: is there enough of value in the existing notes branch to keep it, or is it so much of a cockup that it might be easier to restart from scratch?
18:18:34 <ppawel> I think even if it's better to start from scratch, a quick "lessons learned" document could be helpful for the future
18:18:44 <zere> definitely
18:18:48 <tmcw> yep.
18:19:34 <RichardF> if TomH is happy with the API side of it, I'd suggest perhaps publishing that, with the disclaimer "the UI is b0rked, please consider rewriting that rather than adapting it".
18:21:01 <apmon> Perhaps we could even merge it as a pure API so that external sites and programs can already use it while the UI for gets improved / rewritten
18:21:14 <TomH> well if there was of attaching such a disclaimer in a way that people would see it...
18:21:27 <RichardF> apmon: hey, that's a decent idea
18:21:39 <TomH> but I'll take an action to try and get something pushed out before the next meeting
18:21:45 <zere> ok
18:21:45 <tmcw> Great
18:21:53 <apmon> thanks TomH.
18:21:59 <zere> #action TomH to push current OSB/notes branch public
18:22:03 <TomH> the main open question for the API side of it is what to do about anonymous users - ie whether to allow them
18:22:55 <RichardF> legally it's difficult to take corrections from people who haven't agreed to the CTs
18:23:01 <TomH> the current schema stores both a user name and an optional user id for each update - ie you are logged in then it uses your id and name and if you aren't then you have to supply a name
18:23:08 <zere> #info there are concerns about the UI implementation of the current branch, but it may be possible to merge the API separately so that UI development can proceed independently
18:23:23 <RichardF> i.e. what happens if someone sits there with Google Maps open in another window, and methodically adds bugs for "this road is called n"
18:23:34 <pnorman> anonymous as in unregistered? I'd bounce that off the LWG, I can see problems with that
18:24:54 <zere> perhaps limit the information that you can enter for an anonymous user. to a pre-defined set of options like "this place has closed", "this road is wrong" (in an unspecified manner)
18:24:56 <apmon> If we can make signup even easier than today, it excluding unregistered commenters would be less of an issue.
18:25:14 <apmon> but that is a separate thing to consider
18:25:17 <ppawel> apmon, +1
18:25:26 <zere> indeed.
18:25:29 <ppawel> "login with facebook" :P
18:26:00 <zere> i'm concious that time is running on...
18:26:20 <zere> so, if there are no objections, shall we move on to PL2 tutorial mode?
18:26:24 <zere> RichardF: anything new here?
18:26:57 <RichardF> not especially. I think this is perhaps not a Top Ten Task right now, given the likelihood that whatever beginners' editor solution we have in a year's time is unlikely to be Flash-based.
18:27:09 <RichardF> so there's not a whole lot of point anyone spending time wrestling with Flex on this.
18:27:45 <zere> ok, so you're proposing closing it as unnecessary?
18:27:49 <RichardF> I think so
18:28:00 <zere> get's my vote. how about everyone else?
18:28:08 <tmcw> mine too.
18:28:11 <apmon> perhaps rather then closing it put it as low priority back log?
18:28:30 <zere> oh, sure. by "closed" i meant "not a top ten task"
18:29:21 <apmon> seems reasonable if there is unlikely to be progress on it in the foreseeable future.
18:29:49 <zere> any votes for keeping it?
18:30:24 <zere> ok then
18:31:11 <shaunmcdonald> Maybe turning into a generic tutorial mode of a OSM editor, or is that not specific enough for the top ten list?
18:31:31 <zere> #agreed PL2 tutorial mode doesn't seem like a priority, given the expected move away from Flash as a technology, so we'll move it out of the TTTs
18:31:46 <zere> i thought JOSM did have a tutorial mode?
18:32:13 <shaunmcdonald> Not that I'm aware of
18:33:06 <zere> ok, i'll have a hack at the wording to open it up a bit more. i think a beginner-friendly tutorial mode is wanted, but we seem to be lacking the editor we want to do it in ;-)
18:33:26 <shaunmcdonald> yup
18:33:31 <RichardF> zere: well, that might be a top-ten task in itself... :)
18:33:59 <zere> we'll be coming up on a year soon. don't worry. we'll get a new set of TTTs for 2013.
18:34:30 <zere> next item on the TTTs is the support for multiple languages on
18:34:51 <zere> i don't think there's been any activity on this recently...
18:35:57 <apmon> It also seems that there hasn't really been any progress on this upstream at OSQA
18:35:58 <TomH> I think ris basically declared OSQA insane and gave up...
18:37:01 <tmcw> So, switch from OSQA to something else, or reassign?
18:37:26 <zere> is this still causing problems? being an anglophone, i can't really tell. but i can assume it's underutilised by non-english-speaking communities
18:38:36 <alexbarth> non-english interfaces pose for sure a major hurdle.
18:38:37 <ppawel> I think it depends on community... in Poland people mostly use forums, rarely the mailing list
18:38:56 <apmon> I haven't heard anyone complain about it recently, but I would guess that less people use it than could if it had better multi-lingual support
18:39:53 <apmon> language definitely is an issue, and many contributors seem uncomfortable to communicate in a language other than their native language
18:39:55 <ppawel> and yes, language is a hurdle for some of the Polish mappers
18:40:30 <zere> would it be possible to work around the issue by running multiple OSQA instances?
18:41:12 <zere> i know that's not ideal and may lead to duplicated questions in different instances. but if we're trying to help people in their native language...?
18:41:24 <ppawel> I imagine that's a sysadmin's nightmare.. instance per language
18:41:54 <apmon> It might be worth an experiment to offer it in the couple of main alternative languages. E.g. Russian, Spanish  and German.
18:41:57 <shaunmcdonald> The only way to get around duplicated questions in each language is to translate all of the questions and answers
18:41:59 * alexbarth agreed
18:42:15 <TomH> TBH I don't thnk this is a well solved problem in general - after all StackExchange is still all english
18:42:21 <alexbarth> um, agreed with apmon "sysadmin nightmare … instnace per lang."
18:43:15 <apmon> TomH: Do you have an estimate of how much of a nightmare for sysadmins it would be to have 4 or 5 different language instances?
18:44:06 <apmon> Imho we have a big enough community for the main languages that duplication of questions is not an issue. We currently probably have more helpers than question askers.
18:44:17 <TomH> well I don't think that's a major problem, but it voids the whole purpose of help by fragmenting the community and the questions and answers
18:44:54 <TomH> you'll get the german community with one "correct" answer and the english community with a completely different "correct" answer
18:44:59 <RichardF> "About the multiple site, yes is possible, we currently have 5 sites running over the same install, including this one, but is still not possible to share a database, but it will be in a near future I promise"
18:45:19 <zere> aren't we assuming they're mostly fragmented by language anyway?
18:46:04 <zere> the closest we can come to solving this seems to be fixing the insanity in OSQA/django. but no one seems to relish that task... there might be alternative bits of software, but then we've the fun adventure of migrating...
18:47:57 <shaunmcdonald> There's not much you can do about the language fragmentation, without doing an EU or UN where everything is translated into every other language.
18:48:12 <ppawel> I think it could be better to put the effort into migration rather than going deeper into unmaintained software
18:48:33 * zere stumbles across
18:48:54 <alexbarth> is it correct that interface translation is supported by OSQA?
18:49:00 <alexbarth> imo, that would go a long way
18:49:05 <apmon> zere: Yes, I asked that question when we first added this to the top ten task
18:49:21 <apmon> but there wasn't really much of a useful response from OSQA
18:49:22 <TomH> alexbarth: in theory yes, but ris had trouble getting it working
18:49:42 <ppawel> alexbarth, I've been browsing their JIRA and there are some tickets "this and this is not localizable"
18:49:51 <ppawel> so my guess that it's possible but it's not a pure translation effort..
18:49:56 <ppawel> also fixing the code
18:50:00 <alexbarth> understood
18:50:23 <apmon> My guess would be that unless we go the multiple language instance, this task is also likely not to be completed any time soon?
18:50:39 <zere> iirc, ris had said that somewhere in OSQA it was necessary to choose a single language. and that it was baked-in in some hard-to-change manner...
18:51:06 <zere> so unless anyone's stepping up to do some major python-fu, it looks like apmon is right.
18:51:39 <alexbarth> implementation aside, a very light implementation w/ interface translation and tag-like language identification is by far preferable over more heavy handed implementations like e. g.  running multiple instances.
18:52:01 <alexbarth> so given that no one here seems to be championing this task, does it go off hte top ten list?
18:52:17 <alexbarth> or into an 'important but orphan' list?
18:52:36 <zere> depends on whether we think it's unimportant, rather than unchampioned
18:52:53 <pnorman> I think it's important but annoyingly difficult
18:52:53 <ppawel> to be honest I kind of question the role of vs forums.. perhaps better forum integration instead of trying to save would be better long term
18:53:12 <alexbarth> forums?
18:53:17 <ppawel> but I'm not 100% sure what was the vision behind
18:53:24 <apmon> Imho functionally it is still important. So I would favour the "important but orphaned" category
18:53:32 <RichardF> alexbarth:
18:53:32 <ppawel>
18:53:47 <RichardF> mostly used by people east of the Rhine :)
18:53:55 <alexbarth> hm :)
18:54:08 <zere> apmon: moving it out of the main category and putting it into any other category is basically making it less important
18:54:09 <ppawel> yeah and that's another problem.. fragmentation of people between different channels
18:54:22 <RichardF> ppawel: that's another insoluble problem ;)
18:54:32 * alexbarth has a special dark place in his heart for forums
18:55:14 <apmon> ppawel: The idea was afaik for questions that actually have an answer. In order not to have to read through 150 pages of discussions and interpret things to figure out what the answere= is
18:55:57 <ppawel> apmon, that convinces me... indeed stackoverflow vs forums are very different...
18:56:05 <alexbarth> apmon: and i feel does this well in principle
18:56:21 <alexbarth> it's like stackoverflow fixed forums
18:56:30 <zere> indeed. the up/down voting system is supposed to bring to the top consensus answers, rather than just the first reply.
18:57:02 <zere> the other option on the TTTs item is this shapado thing. anyone know anything about that?
18:57:27 <zere> looks like it's a rails app, so probably already i18nable or close to... well... i'm making a big assumption ther
18:57:46 <zere> fyi:
18:57:47 <ppawel> looks like a cloud solution to me?
18:57:58 <zere> it's open source, so we could self-host
18:57:59 <ppawel> ah, I thought you meant
18:58:02 <ppawel> ok
18:58:49 <ppawel> certainly +1 for rails..
19:00:16 <zere> well, this is showing me a language filter, so it looks pretty good:
19:00:39 <zere> anyone feel like taking an action to look at how feasible an OSQA->shapado migration would be?
19:01:20 <ppawel> hmm, shapado uses mongodb
19:01:25 <ppawel> migration should be fun :)
19:01:42 <ppawel> relational db -> nosql db  :)
19:02:02 <zere> and potentially a very different model of karma, etc... too :-)
19:02:12 <pnorman> ppawel: ya, but osqa is popular enough it's likely a solved problem
19:02:38 <ppawel> pnorman, you mean the migration?
19:03:00 <pnorman> Ya
19:03:40 <ppawel> I can volunteer to take a look at migration feasibility
19:03:47 <zere> excellent, thanks :-)
19:04:20 <zere> #action ppawel to look at feasibility of migrating OSQA to shapado (or some other OSQA alternative)
19:04:38 <zere> it also looks like we're out of time for today
19:04:55 <ppawel> :(
19:04:59 <zere> apologies - i thought we'd get through all the TTTs in one session, but i guess we'll have to continue next week
19:05:01 <apmon> Do meetings have to be only 1 hour?
19:05:21 <ppawel> I really wanted to reach "OWL-powered activity/history tab" to plug my project :)
19:05:43 <zere> ppawel: whodidit?
19:05:57 <ppawel> no,
19:06:11 <ppawel> I wanted to learn about OWL status
19:06:32 <ppawel> it could be potentially plugged into the activity server, if not, I can solve the activity/history problem otherwise
19:06:56 <ppawel> in any case, I'm working in these areas right now
19:07:12 <ppawel> but this is OT, sorry for disrupting the meeting
19:07:37 <zere> yeah... that's on me. there's a bunch of problems with the current code. if you've got time then i'd love to talk to you about it over email or something.
19:08:01 <ppawel> zere, great, will do
19:08:14 <zere> apmon: they don't have to be - we usually overrun any any case... but i think it helps with people scheduling stuff if there's an end time
19:08:37 <alexbarth> LOL
19:08:51 <zere> huh... and the meetbot has also crashed... so much for that one...
19:08:58 <apmon> fair enough. So we will continue with it next week?
19:08:59 <alexbarth> ok, when should we continue? zere - want to just send out a new date?
19:09:22 <zere> same time next monday, if that's ok for everyone?
19:09:24 <alexbarth> would be also cool if point people on TTT tasks would just send quick updates ahead of time
19:09:27 <alexbarth> ok by me
19:09:35 <ppawel> ok
19:09:50 <alexbarth> i'd love to also chat about how the knight grant will play into things. tom and i will post some time this week
19:09:53 <alexbarth> more details
19:10:10 <zere> #endmeeting