Working Group Minutes/EWG 2011-11-28


IRC nick Real name
apmon Kai Krueger
RichardF Richard Fairhurst
TomH Tom Hughes
zere Matt Amos


  • Further work on ensuring that the scope for the tasks on Top Ten Tasks is clear.
    • OSB / notes front-end requirements.
    • i18n for - OSQA doesn't seem set up for translation collaborations, so leave it to ris's judgement.
  • Contact people were added to the top ten tasks page, so that potential contributors would know where to go to get help.
  • Some discussion about routing - hardware needs are still unclear.
  • It was planned to discuss incentives, but ran out of time, so set for next meeting.


18:00 < zere> minutes of the last meeting
18:00 < zere> usual procedure - pls have a look and let me know if there's anything that needs correcting
18:01 < TomH> blimey, forgot about EWG
18:02 < RichardF> (is that the B-side to "Forgot about DRE"?)
18:02 < apmon> Minutes look good to me
18:02 < zere> agenda for this meeting is to talk about motivation/incentives/bounties, look at the top ten again - see if we're approximately totally happy with it, and any other stuff from the F2F.
18:02 < zere> RichardF: "EWG's dead, they're locked in AWG's basement"?
18:03  * RichardF laughs
18:03 < RichardF> and discuss anything that came out of the hack weekend, too?
18:04 < zere> sure
18:06 < apmon> Shll we go through the tasks first, and then talk about the incentives?
18:06 < zere> yeah, ok.
18:06 < zere> so starting with
18:07 < zere> are we happy with the wording of that? anything vague needs to be firmed up? 
18:07 < TomH> no change on that one - got distracted with other things
18:07 < apmon> I suspect that one would need fairly tight coordination with TomH?
18:08 < zere> can we be more specific about what "updated and polished" or "reviewing and cleaning" means?
18:08 < RichardF> I'm happy to help work on the frontend stuff in conjunction with the redesign if needs be.
18:08 < zere> i suspect all of them need fairly tight coordination with the OSM dev community in general.
18:08 < apmon> We probably should also add a contact person to each task
18:10 < TomH> the javascript basically needs rewriting to be more light weight and move more of the rendering to the server
18:10 < TomH> because trying to do anything with the html and it's styling is impossible while it's being generated in a zillion lines of jscript
18:10 < TomH> ajax FTW
18:12 < zere> what do you mean? when someone clicks a bug, for it to get the details from the server as an html snippet?
18:13 < apmon> I'd prefer to keep as much client side as possible to not introduce unnecessary latency
18:14 < TomH> zere: well it does a call at the moment that returns ja that is just a long list of calls to an already loaded client function
18:15 < TomH> that function that does a zillion createElement etc calls and is a completely unreadable and unmaintable mess
18:15 < TomH> I want to make it the return include preformatted HTML for the popups rather than building the HTML in the client
18:16 < TomH> it can still be one call that returns all the data
18:16 < zere> surely that increases the size of the returned data by a significant factor, doesn't it?
18:16 < TomH> a bit sure, but is that a major issue
18:17 < TomH> the current code really is unreadable, and the HTML it generates is not nice looking and I have no idea how to change it without going mad
18:18 < zere> how about something like ?
18:18 < TomH> personally I'd rather have something that looks vaguely like HTML that I can edit
18:18 < TomH> that would probably work as well
18:18  * RichardF suggests innerHTML and runs away very fast
18:19 < TomH> RichardF: well sure, how do you think I plan to inject the HTML the server gives me...
18:19 < RichardF> :)
18:20 < TomH> I mean yes you could write js that builtup a big string of HTML but I'm not sure it would be any more readable
18:20 < TomH> well maybe a little
18:21 < apmon> TomH: Are there other areas you have identified that need rewriting?
18:22 < TomH> I think the server is mostly sorted, apart from hopefully getting rid of the hacky code that returns the js script
18:22 < apmon> I can update the wiki page to make it more clear where the clean up needs to happen.
18:22 < TomH> hopefully we either do a better version of that which returns something with HTML in or we just make the frontend use the geojson or something
18:23 < zere> apmon: i have added some notes - feel free to edit.
18:23 < apmon> And possibly look at it my self to see if I can improve on in that area
18:23 < zere> and i've added both apmon and TomH as contacts for this.
18:23 < apmon> sounds good
18:24 < zere> next item is PL2 tutorial mode. RichardF, do you have anything to add to 
18:24 < zere> and can i add you as the contact, or gravitystorm?
18:24 < RichardF> add me as the contact by all means. nothing to add to it.
18:25 < TomH> sorry, stupid laptop decided to hibernate...
18:26 < apmon> Next item then?
18:29 < zere> yup, sorry.
18:29 < zere> i18n for PL2. was that done completely at the hack weekend?
18:30 < RichardF> we hope so, yes. am planning to move it into the production P2 later today, hopefully.
18:30 < apmon> Sweet
18:30 < RichardF> obviously people still need to actually translate stuff, but that's more of an ongoing thing anyway.
18:30 < apmon> Can it be added to translatewiki then?
18:31 < RichardF> would need to liaise with gravitystorm about that - potentially yes but I think we need to make sure that the i18n code is working consistently first.
18:32 < zere> next item is support for i18n on
18:33 < zere> i know ris did some work on this at the hack weekend
18:33 < zere> anyone here know python/django?
18:33 < TomH> yes we have enabled it, but I don't think there are many translations
18:33 < TomH> the german one was completely empty so we had to remove it because it was crashing django
18:33 < TomH> the french one has a few messages translated but not many
18:34 < zere> so do we need to announce it and ask for translators?
18:34 < TomH> well that's an uptream issue, we need to be careful about how much we put ourselves in the middle
18:35 < apmon> My impression was from looking through that most people translate their own templates and not too much sharing is going on upstream :-S
18:35 < zere> ok... announce it and heavily hint that anyone whose language doesn't have many translated strings could help over there ----> ?
18:36 < TomH> probably best to ask ris what he thinks the beast appraoch is
18:36 < apmon> OSQA doesn't really have a system to collaborate on translations at the moment it seems
18:36 < zere> ok, let's leave it to ris. any idea what his wiki username is?
18:37 < RichardF> Robert
18:37 < apmon> So perhaps we can do them in our repository and then once we have complete translations offer it back upstream
18:38 < zere> next item is basically empty. so what did we all want last week when we wanted ODbL migration tools?
18:38 < apmon> The second part of the task is the question of how to deal with different language questions and answeres
18:38 < RichardF> zere: well, one tool is the OSMI view which Frederik is working on :)
18:39 < zere> so viewing what data will need to be removed.
18:39 < zere> anything else?
18:40 < TomH> apmon: the messages are part of the upstream svn repo that we are using, so we don't really have a way to separate them out to a local repo
18:43 < apmon> Can we ass Frederik as the contact for the ODBL task then?
18:44 < apmon> *add*
18:44 < zere> i am doing just that ;-)
18:44 < zere> and addign some more descriptive text.
18:45 < zere> argh. what's that map that shows ODbL clean-ness in red/yellow/green?
18:47 < zere> ko
18:47 < zere> ok
18:47 < zere> next: clickable POIs on the front page
18:48 < zere> ok....
18:48 < zere> next: routing frontend?
18:49 < zere> ok....
18:49 < apmon> with regard to clickable POIs, we might be able to link to www.ö as an example
18:49 < apmon> or
18:49 < TomH> I rebased the routing branch on top of the current trunk with rails3+jquery on saturday
18:50 < apmon> Thanks, I will have a closer look at it again to see what more needs to be done
18:50 < zere> apmon: except that's not working at the moment, apparently...
18:51 < apmon> But there is still quite a bit to do to get it into a user friendly presentation
18:51 < apmon> zere: Ah, yes OpenLinkMap doesn't seem to work
18:51 < apmon> OePNVkarte though those (at least partially)
18:51 < zere> apmon: have assigned you as the contact for routing. hope you don't mind?
18:51 < apmon> Yes, that's fine
18:53 < zere> next is the XML deleted items call. i'm happy to be contact for that, if no-one else does?
18:55 < apmon> I presume you will also be the contact on the OWL task?
18:55 < zere> that probably makes sense. maybe someone else wants to take the deleted items call?
18:56 < zere> anyone?
18:57 < zere> ok... moving on. anyone want to be the point of contact for the routing backend?
18:58 < zere> apmon: you're contact for the frontend. does it make sense to be the contact for backend too?
18:59 < apmon> I guess so. 
18:59 < zere> :-)
18:59 < apmon> I am not really familiar with any of the code, but I can help with evaluating resources and such
18:59 < apmon> Did you get a better feeling for what is needed for osrm over the weekend?
18:59 < zere> i think that's what's required at this point
19:00 < zere> yeah. i got it compiled and importing on dev. dennis was saying that he thinks it's much less resource-intensive than it used to be.
19:00 < zere> but we're still talking about an impressive amount of memory / cpu to update the graph.
19:01 < apmon> Do we still need gosmore for the more diverse routing modes?
19:01 < zere> and we still need to know whether running something like gosmore would actually be better-suited to our case.
19:02 < apmon> Is there a budget for the server hardware?
19:02 < zere> to be honest, i don't think we need diverse routing modes. just car & walking would be fine by me. many would want cycling too.
19:02 < zere> no
19:02 < apmon> zere: OSM is all about the diversity... ;-)
19:02 < zere> this exercise is to find out what kind of hardware might, maybe, at some future point, be necessary
19:02 < zere> apmon: that's why we distribute our data, yes ;-)
19:03 < apmon> If there is a funding drive soon, should the potential for the server not be included?
19:03 < zere> no
19:04 < apmon> why not?
19:04 < zere> when this task is finished we'll know what resources are necessary and, if the routing front-end is also finished, then it's time to consider maybe buying a server
19:04 < zere> because at the moment, "what server do we need" is the same as "how long is a piece of string"
19:05 < zere> all we know is that routing is *hard* and might need a big server.
19:05 < zere> but that's no reason to budget a smaug-class machine, just on the off-chance that this work eventually gets done.
19:07 < zere> i guess we'll have to leave the discussion about motivation / incentives / bounties for next time.
19:07 < zere> anyone have anything else to discuss?
19:08 < apmon> Are we going to announce the top ten tasks?
19:08 < apmon> We can probably gauge the reaction to the list before thinking about insentives
19:08 < zere> yes, i guess that would be a good idea :-)
19:08 < RichardF> (sorry, have to disappear - back in about 15 mins if meeting still going on)
19:08 < zere> RichardF: worth doing this via CWG, or straight to dev@?
19:10 < apmon> probably straight to dev would be fine
19:12 < zere> ok. shall we postpone that to next meeting as well?
19:16 < zere> i guess so.
19:16 < zere> thankyou to everyone for coming! i hope to see you next week :-)