Working Group Minutes/EWG 2015-02-02

Attendees

IRC nick Real name
pnorman Paul Norman
RichardF Richard Fairhurst
RobJN Rob Nickerson
tmcw Tom MacWright
zere Matt Amos

Summary

  • Mobile development
    • RobJN has written up some notes on a potential mobile app to make it easier for people to collect survey data PhotosAndNotes.pdf.
    • RichardF had a similar-sounding idea 2 weeks ago, and has also written up some notes [1].
    • One thread of the discussion was about getting "official" status for the app(s), so the developer would appear as "OpenStreetMap".
    • ACTION: RobJN to take something forward to the board broaching the topic of having an official class of apps.
    • ACTION: zere to start organising a hack weekend with mobile dev as the theme, provisionally for April.
    • tmcw pointed out coffeedex as an example of special-purpose editors.
  • Raspberry Pi
    • RobJN noted that the new RPi2 has been released, and wonders whether anyone has any ideas for how it could be used for something OSM-y.


IRC Log

17:33:12 <zere> minutes of the last meeting: http://wiki.osmfoundation.org/wiki/Working_Group_Minutes/EWG_2015-01-26 - please let me know if anything needs changing.
17:33:38 <zere> on the agenda today is: mobile dev.
17:33:51 <zere> #topic mobile dev
17:34:35 <RichardF> good evening
17:34:49 <RobJN> evening
17:35:57 <pnorman> Could someone introduce the topic?
17:36:16 <zere> pnorman: please wait, organising stuff
17:36:21 <pnorman> k
17:37:05 <RobJN> So I guess the question is about next steps on app ideas (both Richard's and if there is interest in it, mine)
17:37:35 <zere> RobJN (Rob Nickerson i assume?) has written up some notes on a potential mobile app to make it easier for people to collect survey data. he's agreed (i hope) that we can share it and i've uploaded it here: http://wiki.osmfoundation.org/wiki/File:PhotosAndNotes.pdf
17:38:23 <zere> RichardF had a similar-sounding idea 2 weeks ago, and has also written up some notes: http://www.systemed.net/osm/osm_mobile.pdf
17:38:37 <zere> (i think in response to RobJN's notes)
17:39:50 <zere> so... now we have some documents setting out the idea - the questions are: 1) are they potentially the same app? 2) how do we help kick off development of it/them?
17:39:50 <RobJN> Yeah happy for you to share that :-)
17:41:21 <zere> i don't know if RobJN and RichardF would agree with me here, but based on the two documents, i'd say RichardF's is considerably more complex as it involves editing in addition to data collection.
17:41:41 <RichardF> yep, I think that's fair.
17:41:57 <RobJN> Re 1) - Possibly, yes, it just depends on whether you want a really simple app with no editing capabilities, plus a simple app with basic editing.
17:42:10 <zere> but does that mean a P2-style "expert mode" in the same app, or does the inclusion of such a mode spoil the UX of either or both?
17:42:33 <pnorman> I'd say RichardF's is a more complex app but RobJN's involves more backend changes
17:42:39 <RichardF> I'd envisage Rob's idea as something that could potentially be developed by the community in a relatively short space of time, particularly if you start off Android-only (there are more likely to be interested Android developers in and around the existing community).
17:42:58 <RobJN> I think that given that the basic bits need doing, perhaps we start with this, release something, and then start adding features until we get to Richard's idea
17:43:13 <RichardF> that's then a potential stepping stone to a consumer-grade app along the lines which I've mocked up, and where we'd need to look crossplatform (i.e. iOS too)
17:43:34 <RobJN> We end up with two apps but this is fine (we have multiple editors catering for different needs)
17:43:49 <RichardF> and a lot of the backend work for Rob's app would be invaluable for a consumer app, particularly the pic/audio storage
17:44:18 <RobJN> Exactly :-)
17:45:37 <RobJN> Although I think we should aim to do iOS and Android from the off. Not sure what developers think of this. I've played with Appery.io before which can spit out both Android and iOS versions. Seemed quite good but I strated to get stuck with the more technical bits.
17:46:09 <RichardF> yeah - the problem with using a crossplatform solution like that is that you're less likely to find devs who are experienced in it
17:46:41 <zere> the mobile devs that i know all pour scorn on "cross platform", but i don't know how much of that is bigotry and how much is really due to benefits of being platform-specific.
17:46:46 <RichardF> I suspect we have <reasonable number> of Android devs in and around the community, <small number> of iOS devs, and <almost zero> of any crossplatform solution
17:47:07 <RichardF> I mean, I could write something in Flash that uses Adobe's AIR crossplatform compilation stuff, but it might not be very popular ;)
17:48:08 <zere> from the start, though, i'd like to kill the idea of "backend changes". we shouldn't be putting anything extra into rails_port for this. if we need to write new, independent services; okay. but at a time when we're trying to break up rails_port to make it more maintainable, i don't want anyone to get the idea that we'll be adding to it.
17:48:37 <RichardF> zere: +1
17:49:12 <RobJN> One of the things I would like is to open an OpenStreetMap portal on the main app stores (to go hand in hand with openstreetmap.org). Do you think we can have some apps as android only or should we be aiming for cross-platform?
17:49:45 <RichardF> depends on the audience, really. existing mappers are likely to skew towards Android.
17:49:52 <RichardF> (I say that as an iPhone owner :) )
17:50:17 <zere> having never released a mobile app, i have no idea what a portal in an app store would mean. any sort of "official" presence is likely to require bureaucratic overhead.
17:50:39 <RobJN> So we would need some way of storing collected photos online. The natural thing is to have this and Notes together, but if it's easier to keep them seperate to start then that's fine too.
17:51:12 <RobJN> I mean having the developer listed as simply "OpenStreetMap"
17:51:25 <zere> RobJN: for technical reasons, it's easier to separate them. however, the user of the app doesn't need to be aware of that split - it would transparent to them.
17:51:40 <RichardF> photos are essentially an exercise in cloud storage. it doesn't have to be (shouldn't be) anything to do with OSM core infrastructure.
17:51:46 <RobJN> zere: correct.
17:52:28 <zere> right, but having the developer listed as "OpenStreetMap" means that OSMF will be the owner, will take any liability, etc... that's going to require a lot more hoop-jumping than simply releasing an app.
17:52:38 <RobJN> But it would be nice to have some level of "officialness" just like the defualt editor on osm.org
17:52:53 <RobJN> So making the apps available at: https://play.google.com/store/apps/developer?id=OpenStreetMap
17:53:10 <zere> then there'll be continued effort put into 1) preventing other people from using the OSM brand, 2) evaluating other apps which wish to be under the OSM portal/umbrella
17:53:11 <RobJN> zere: correct, but let me have a go :-)
17:53:21 <RichardF> zere: to be fair, we do 1 already
17:53:55 <pnorman> We should probably see if we can start jumping through the hoops, if solely to stop others from claiming their apps as official when they aren't
17:54:01 <RobJN> I can approach the OSMF with the idea and they can knock it back if they want, but don't ask=don't get
17:54:03 <zere> we do (1) very badly at present, partly, i think, because we don't have an "official" app and partly because no-one puts in the continued effort to do it
17:54:31 <zere> oh, i wasn't suggesting we don't do it. i was simply "managing expectations" ;-)
17:56:07 <RobJN> So the bit I want to add to my report before sending it to OSMF board is some endorsements basically stating that you are happy to see the app on the play store under OpenStreetMap in theory, reliant on the code being of sufficient quality.
17:56:58 <zere> i think that's premature. there's no point writing any sort of report or asking for any kind of permission until there's an actual app on an actual store.
17:57:22 <pnorman> I would like to see the EWG recommend OpenStreetMap claimed on the stores independent of the app
17:57:41 <RobJN> Well I was hoping that it might spark some interest in it so that I can find developers will to work on it :-)
17:57:48 <pnorman> Although, do we need an app to open an OpenStreetMap portal?
17:58:56 <RobJN> Basically I've taken it as far as I can, so the main way I can help now is by doing that "hoop jumping" (both for myself and Richard)
18:00:36 <zere> what i'm afraid of is that you take a proposal to the board, they discuss it, hopefully approve it, and... then nothing happens. a year from now, someone comes along with a working app and is now in a worse position because they're "blocked" by a vapourware app which never existed.
18:01:14 <RobJN> So I was thinking of a time limit. Say 1 year.
18:01:18 <zere> additionally, board approval of a "sight unseen" app just makes it a complete wildcard. people will rush to put anything on that pedestal, regardless of the actual quality.
18:01:39 <pnorman> I wouldn't want to recommend anything specific to this app. Should we say an app of sufficient quality, with potentially other criteria put upon it?
18:02:37 <RobJN> Really I'd be asking the board to agree to the potential for OpenStreetMap to be used as a name on the app stores and to establish some acceptance criteria
18:03:16 <RobJN> The app would be an example of what we could put on the store
18:03:36 <RobJN> Richard's suggestion would be a second example.
18:04:01 <RobJN> And if you wanted a full blown editor then there are Android and iOS examples already in existance.
18:04:14 <pnorman> I agree with recommending for the potential for the name to be used, but not with tieing it to any app yet. I think it's premature to have acceptance criteria - also, keep in mind we didn't set down acceptance criteria for iD before it was developed
18:05:18 <RobJN> I just don't want to see the goal posts moving every time!
18:05:44 <zere> right, so would it be better to talk to the developers of those existing full-blown editors (Vespucci, i guess?) and see if they would want to use the OSM name and accept whatever criteria that come with that?
18:06:19 <RobJN> Could do. Question is do we want the first OpenStreetMap app to be a complex one.
18:06:40 <RichardF> well, that depends if anyone writes it ;)
18:06:59 <zere> that ship has rather sailed - vespucci exists (and is complex) :-/
18:07:16 <pnorman> RobJN: I'm not opposed to setting criteria, just setting criteria this early. The criteria may differ depending on the target of the app, what else exists in the space, and other factors
18:07:21 <RobJN> The ship is the "OpenStreetMap" portal on the stores
18:07:22 <zere> whether it's "official" is, i would suggest, not spectacularly important to 99% of users.
18:08:06 <RobJN> Possibly not but it gives people a single place to go to to find OSM apps
18:08:35 <zere> i think people just put "openstreetmap" into the app store search box, don't they?
18:08:40 <pnorman> yes
18:08:47 <RobJN> It also enables you to then promote it via the CWG
18:08:50 <zere> then look at how many stars it has, reviews, screenshots, etc...
18:09:34 <zere> i guess what i'm trying to say is that i think the quality of the app is more important than its official status (not that the latter is unimportant, just less important)
18:10:02 <RobJN> functionality/UI as well
18:10:19 <pnorman> that's part of the quality of the app
18:12:10 <zere> as for getting this written, i can see a few ways forward: 1) we've been thinking about running a "themed" hack weekend, and this might be an ideal candidate. not, perhaps, that it can be finished in a single weekend, but a good start can be made and hopefully momentum kept up afterwards
18:12:11 <RobJN> So given that I'm not a dev, my opions are to (a) throw this idea out on dev ML and see if anyone picks up up out of their own good will, (b) try and get some OSM sign-on first to then hope that someone is more likely to pick it up, failing that start looking for funding/sponsorship (which would be near impossible advertising this as a "Rob Nickerson" app)
18:12:50 <RobJN> A hack weekend sounds great
18:13:54 <zere> 2) take the idea to the board and get *WG (i guess either CWG or EWG) to run a competition: "we're looking for an official app, here's a document with some ideas, here's a panel of judges and judging criteria. the best app on the DD/MM/2015 will be blessed."
18:14:03 <RobJN> (p.s. Vespucci on Android, Go Map!! on iOS were the two I was thinking off)
18:14:33 <RichardF> 1 is good. I'm not sure we're anywhere near 2 yet
18:15:01 <RichardF> and having a lot of developers competing with each other would be a really bad way of doing things
18:15:07 <zere> 3) make a more flashy set of mock-ups and run a crowdfunding campaign to get it written. (to my mind, this is the highest risk and lowest payoff version - we get no continuity of app development, and run a high risk that the project will run out of money and fail)
18:15:25 * RichardF gets back to the Swift course
18:15:35 <zere> RichardF: yes, the downside to (2) is that we'd get competition where collaboration would really be more helpful
18:15:58 <RichardF> I'd suggest we start with 1. the next steps will become more apparent as we actually start to build things
18:16:35 <RobJN> With (2) you could ask devs to approach you before writing any code. See who is interested and then put a small group together??
18:17:00 <RobJN> Was (1) hack weekend? or just to put the idea on dev ML?
18:17:04 <RichardF> hack weekend
18:17:08 <RobJN> ok
18:17:36 <zere> no reason not to do both. personally, i think it makes sense to start with a hack weekend, but there are downsides to that too.
18:17:47 <RobJN> My concern with dev ML is that it opens you up to the wild west, things getting added to and losing the simplicity of the idea
18:18:39 <zere> that's true. it's a quesion of how much "top-down control" you want to exercise and how much you're willing to let that put off developers.
18:18:44 <RichardF> also, I'd gently add in here that SOTM-US is June 6-8 and will (I suspect) have more OSM-inclined mobile developers than any other event has yet had
18:19:11 <RichardF> zere: there are actual developers (the sort that develop) on the dev@ list? :p
18:19:42 <zere> anyone who's volunteering their time (and even a paid dev) is going to want to have input into the idea, which will tend to drive it away from the purity of the original concept.
18:19:51 <pnorman> Regardless, I'd like to get the wheels turning on creating an OpenStreetMap portal and in principle saying that apps may be added to it if they meet yet to be determined criteria
18:19:58 <zere> RichardF: i think so. generally lurking, though ;-)
18:20:54 <zere> pnorman: "yet to be determined criteria" seems synonymous with "continually moving goalposts"
18:21:32 <pnorman> zere: I don't see it as that. An editing app may have a different set of criteria than a data collecting app
18:22:06 <RobJN> zere: well yes I appreciate that, and I would dare suggest how to code it, I do feel that we should agree the design up front and stick to that. It has to be dead simple otherwise it won't avheive it's objective
18:22:15 <RobJN> *wouldn't dare
18:22:33 <zere> pnorman: sure, but if you're not prepared to think about, write down and be transparent about the criteria up-front then it's very confusing and quite possibly scary for anyone attempting to meet those criteria
18:22:42 <RobJN> (long day! and I can't type at the best of times!)
18:23:23 <pnorman> zere: perhaps say something about criteria to be determined in advance of development? but then we've got the possibility of adding existing apps
18:24:49 <RobJN> On criteria you could say that a data collecting app should upload to a central place so that any map editors could use the collected data. The uploading should be direct from the app so no faffing with copying the files to a computer first.
18:25:08 <zere> i think i'd be happy to try and run a hack weekend (perhaps in April?) on the theme of a simple OSM data collection app. i'd like RobJN to come along to explain the concept, and RichardF to sprinkle potlatch-dust on it. sounds good?
18:25:16 <RichardF> I'm up for that
18:25:36 <pnorman> zere: in many ways I see it as someone proposing a project to come up with criteria, along with milestones, etc and all the other standard project proposal stuff
18:25:44 <RobJN> I'm in (assuming we are looking at the UK as a location)
18:26:36 <RobJN> So actions: (1) Hack weekend in April (2)??
18:26:54 <zere> pnorman and RobJN: do you want to take something forward to the board broaching the topic of having an official class of apps? given the board's historical decision-making speed, perhaps the hack weekend will occur first ;-)
18:27:36 <pnorman> I'm not going to have time to do anything before the karlsruhe hack weekend
18:28:08 <zere> pnorman: criteria, milestones and standard project management stuff are (in my experience) in fact actively harmful to open, collaborative, "bazaar" projects.
18:28:08 <RobJN> zere: Can do. pnorman would you have enough time to get me on email and suggest what I should be preparing?
18:28:56 <pnorman> zere: you said we want acceptance criteria
18:29:56 <pnorman> RobJN: have you ever done a project proposal?
18:30:03 <zere> pnorman: if you're running a competition, or setting up an open class - yes.
18:30:30 <zere> s/setting up an open class/asking people to submit their apps for approval/
18:31:33 <zere> milestones and project management imply continued development under OSMF's supervision which... is too horrific to comprehend.
18:31:51 <RobJN> pnorman: I've put projects out to tender and seen the reports that have come back in. I've got the pdf written up that was linked to at the start of the meeting.
18:32:22 <RichardF> I think we have to accept that any "official OSM" mobile app would have a developer-maintainer, just like all other OSM software does.
18:32:37 <RichardF> (TomH for rails_port, me for P2, jfire for iD, and so on.)
18:32:43 <RobJN> I'm not wanting to do any rigorous project management as it's probably too much for an open community
18:32:58 <RichardF> having it ruled by a non-developer or, god forbid, a committee would be a disaster.
18:33:01 <RichardF> let's stick to what we know works.
18:33:39 <RobJN> The main ongoing work is to keep any Photo server up and running
18:33:56 <pnorman> RobJN: nor am I - but if you know how the more rigorous project management works it's helpful as then you can strip away parts that aren't desirable
18:34:04 <RobJN> The app should be so simple that it rarely needs an update (unless iOS / Android radically change)
18:34:06 <zere> RobJN: i think there's a case to be made that an "official" stamp - being included in the OSM apps umbrella - would help. the next step on that is thinking what the minimum criteria are (e.g: no spyware. no ads? works globally? supports at least these translations? etc...)
18:34:15 <RichardF> hmm... photo server... maybe we've finally found something to do with Amazon's offer of support ;)
18:35:13 <RobJN> zere: Yeah really basic criteria like what you have just suggested is what I had in mind.
18:35:14 <zere> RichardF: s/offer of support/desire to maximise PR while minimising actual usefulness to OSM/ sure.
18:37:44 <zere> okay... so i think we have our two sets of actions: i'll start planning for a hack weekend with this as a theme (helped by RobJN and RichardF), and RobJN will continue to make a case for "official" status on the app store and the minimum criteria for that.
18:37:48 <zere> sounds good?
18:38:17 <RobJN> Yep, sounds good to me.
18:38:48 <zere> awesome :-)
18:38:53 <zere> #topic AoB
18:38:59 <tmcw> just fyi, might want to look at coffeedex, this is the problem space it is a demo of
18:39:21 <zere> anyone else have something to discuss (hopefully very quickly, as we're over time already)?
18:40:18 <RichardF> not me
18:40:21 <RichardF> (and coffeedex is fun)
18:41:18 <pnorman> Just noting that I'm hoping to tag a new osm2pgsql minor version (0.87.3?) soon
18:41:22 <zere> it's certainly part of the discussion we'll have to have (native vs. whatever you call a locally-installed web app without the web)
18:41:39 <RobJN> What's possible with the new Rasberry Pi (that was announced today)? Anything interesting that could be done with it as a way of introducing OSM to people?
18:41:41 <tmcw> yep. factor in react-native for that conversation
18:42:03 <zere> yeah, i watched the keynote for that, it looks totally awesome.
18:42:14 <RobJN> (That's my randon off topic asside that I thought of whilst listening to the radio today)
18:42:58 <pnorman> RobJN: I might buy one and set up osm2pgsql on it
18:44:07 <zere> RPi is definitely targetted at education, but the concept of having very cheap phones-but-not-mobile available in developing countries is very interesting.
18:45:17 <RobJN> zere: No maps in education?
18:45:38 <RobJN> OS have a pretty good grip on UK education in that space
18:45:46 <zere> yeah, i guess that didn't make much sense...
18:46:29 <zere> i meant; the RPi specifically seems targetted at teaching kids how computers work and how to program them (i.e: like the BBC micro of yore)
18:47:08 <RobJN> That's the aim of the RPi Foundation. And from what I can see, they seem to be doing a good job so far.
18:47:42 * pnorman is about to order two
18:47:42 <RobJN> Anyway I've got to head off. Thanks for dedicating so much time towards the mobile app ideas.
18:47:49 <zere> however, there's no reason that geography (or graph theory) couldn't be taught using OSM data.
18:48:43 <zere> but they've also sparked an explosion in RPi-like small, cheap devices which have many potential uses outside the classroom.
18:49:30 <pnorman> hah. element14 is down, raspberri pi 2 is released. any relationship?
18:50:22 <zere> okay, thanks to everyone for coming & hope to see you next week! :-)