- zere, emacsen and apmon had attempted rails_port installs. All found it harder than it really should be, generally on account of the large number of steps rather than any inaccuracy of the instructions on the wiki.
- emacsen is working on a VM image, which will reduce the number of steps.
- TomH and SteveSn_ went through trac tickets and closed many which were outdated or otherwise no longer necessary.
- Many of the trac tickets can be thematically grouped.
- There was a discussion of whether trac can support the grouping and categorisation that we'd like for organising "projects".
- All to spend some time going through trac to get a better idea of what keywords we'll need. (Since they have to be declared up-front)
- It was discussed whether developer-oriented videos would be helpful, but the consensus was that written documentation is more useful as the reader can go at their own pace.
- IRC workshops were discussed, but the consensus was that it's a very time-consuming way to transfer knowledge.
- apmon suggested an on-IRC hack weekend, and offered to organise it.
- apmon & zere to present a "Developing OSM 101" lightning talk at SOTM.
- There was a discussion about collaboration opportunities with the also recently set up design mailing list.
18:01 <@zere> just the 6 of us here this time... oh well, maybe some stragglers will join later.
18:02 <@zere> minutes of the last meeting, for your reading pleasure. if there's any objections or innacuracies, please let me know: http://www.osmfoundation.org/wiki/Working_Group_Minutes/EWG_2011-08-22
18:02 < mkl1> howdy
18:04 <@zere> since emacsen and i are here now, we can present what we found trying to install the rails port. 18:04 < emacsen> You want to go first or shall I?
18:04 <@zere> i tried installing on OS X and had an annoying time. largely, i think because i already had a ruby install on this machine and they kept conflicting.
18:05 <@zere> emacsen: i'll keep it short ;-)
18:06 -!- SteveSn_ [~firstname.lastname@example.org] has joined #osm-ewg
18:06 <@zere> basically it seems that it's not so much the rails port itself that's hard to install as much as it's all the dependencies. the least painful way seems to be via homebrew, but that's only one of the many OSX packaging systems available.
18:06 <@zere> also, the XCode install seems to be a requirement, which is painful, since it's a 2GB download.
18:07 <@zere> i wasn't a fan of the VM approach before actually trying this. now i'm leaning towards that - since a minimal VM could probably be produced in less than 2GB!
18:08 <@zere> i can write up the steps in more detail, but i'll do that offline. emacsen, how did you get on?
18:08 < emacsen> I took the approach of a VM image with a fresh Ubuntu 10.10 (the long term stable)
18:09 < apmon> I tried to install the rails_port on OSX too, but only got as far as trying to install postgresql (which surprisingly already failed with a bunch of errors).
18:09 < emacsen> The steps on the wiki are complete, but a bit confusing. There's a lot of "go do this stuff and go back"
18:09 < emacsen> the depenedencies use packages, third party packages, gems, etc.
18:09 < emacsen> I tried a 2gb install. That was too small. I'm working through a 4gb install now
18:10 < emacsen> I think there are too many steps for a regular person to go through, so I think the VM image is really going to end up being the only easy choice, but we can also clean up the wiki and make the process more straightforward by reducing the number of "optional steps"
18:10 < emacsen> (done)
18:10 <@zere> emacsen: awesome, thanks :-)
18:11 <@zere> SteveSn_: how did you get on with the trac cleanup?
18:11 < TomH> is somebody that isn't capable of doing it going to be capable of writing reasonable code though?
18:12 < apmon> TomH: I think it is not necessarily "capable" the problem, but "bothered"
18:12 < TomH> that may be a better point
18:12 <@zere> i don't think it's just about the code. there might be people who could help with the css and views stuff that's not really "code".
18:12 < SteveSn_> I went through a handful of old oudated bugs and closed them, mostly under admin and utils. The rails_port has a lot of bugs and I haven't read all of them, but with a small sampling I didn't find a lot that were obvious too close. I hadn't looked at any of the mapnik or potlash etc. bugs
18:13 < TomH> I went through all the admin, api and website bugs
18:13 < TomH> Closed several dozen
18:13 < TomH> and added comment to some and cleaned up titles, set some to enhancement instead of defect etc
18:14 < apmon> Only indirectly related, but is there a way to see the most recent activity on trac tickets? The standard view only has a sort option of date created, rather than date modified
18:14 < TomH> well the timeline will show changes
18:15 <@zere> total, there's almost 1,100 "issues" in trac. what proportion of those do you think could be closed or merged? are so many of them unique and useful?
18:16 < TomH> well in the categories I looked at, which are the web site related ones, I closed all the obvious dupes and already fixed bugs
18:16 < SteveSn_> apmon: If you add the "Modified" column under "View tickets" custom query you can then sort by it
18:16 < TomH> some may turn out to be already fixed with more investigation
18:16 < TomH> most of what is left are enhancements, and I mostly didn't close them unless they were really silly
18:17 <@zere> cool. thanks TomH & SteveSn_.
18:17 <@zere> anyone volunteering to continue to trawl through trac?
18:17 <@zere> and any candidates for small "starter" projects?
18:17 < TomH> there are broad themes sometimes where there are a group of tickets that don't ask for exactly the same enhancement but tackle the same broad area like "tickets about making the nearby user list better!"
18:18 <@zere> any way that we can group them easily in trac?
18:18 < TomH> not off the top of my head - don't think it has dependencies like bugzilla
18:18 <@zere> does it have "project" features, maybe? or pull them onto the wiki for categorisation?
18:18 < TomH> no idea, and there may be things we can do if we add a plugin
18:18 < emacsen> keywords?
18:18 < apmon> I have seen a couple of small "starter" projects which were mostly related to the "social" features of the site
18:19 <@zere> does the "milestone" feature offer anything useful?
18:20 < TomH> maybe, but I suspect it's not very helpful when trac contains so many diverse projects
18:20 < TomH> because milestones are AFAIK global
18:20 < SteveSn_> can we overload keywords for this, ie add 'social' to the keywords then have a report that pulls anything with the given keyword
18:21 <@zere> ok. shall we try adding keywords, like emacsen suggested?
18:21 < emacsen> I am checking... I thought trac had keywords but I don't see this...
18:21 < emacsen> oh there it is. sorry
18:21 <@zere> i propose that all of us take some time in the next week to go through some unkeyworded tickets and add them. sound good? +1?
18:22 < TomH> there are at least three plugins to add some sort of dependency tree
18:22 < SteveSn_> zere: it would be good if we had a place to agree on common keywords so we use the same keyword of the same type of thing
18:22 < TomH> and a tag plugin
18:23 < emacsen> how are tags different than the builtin keywords?
18:23 < TomH> http://trac-hacks.org/wiki/TagsPlugin
18:23 < emacsen> they're easier :)
18:23 < emacsen> gotcha
18:23 < TomH> oh in fact it uses keywords as tags
18:23 < TomH> but it adds tags for wiki pages
18:24 <@zere> install that, then?
18:24 < emacsen> is there a way in trac to see a high level overview? like bugs per component, etc?
18:24 < TomH> well not clear if we need it
18:24 < emacsen> because right now it's hard to see the forest from the trees
18:25 < TomH> http://trac-hacks.org/wiki/KeywordSuggestPlugin provides auto complete for keywords which may be more useful
18:27 <@zere> SteveSn_: what do you think? social has already been suggested. what other keywords can we think of?
18:27 < emacsen> Not to be a jackass and go meta, but what is the end goal of this excercise of working on the trac? To close tickets? to make it easier to see tickets? What is our measurable goal?
18:28 <@zere> i think the goal is to see if we can make it to work as a system for keeping this information.
18:28 < SteveSn_> zere: without going through trac with that in mind I can't say (I don't know the themes of the bugs well enough). I'm thinking if as people started adding keywords they also add that keyword to a wiki page so we can see the 'keywords so far' before creating a new one
18:29 <@zere> so, yes, to close and consolidate tickets so that it's easier to find information.
18:29 < emacsen> I'm reading trac tickets that are years old and thinking "This will never be solved"- shall I just close tickets? If I see easy to fix rendering tickets, send patches? I guess I'd like to know what it is we're using to measure our succcess
18:29 <@zere> to organise the tickets so that someone with a particular area of interest can find what they want to work on.
18:30 < emacsen> okay, so the goal is to make it easier to see. And to that end, we're changing the visuals, and closing tickets. Correct?
18:30 < TomH> SteveSn_: that's why I suggested the auto-complete plugin so that you can more easily find tags that already exist
18:30 < SteveSn_> TomH: that makes sense
18:30 <@zere> SteveSn_: we could catalogue them on the wiki linked to the trac search for that keyword.
18:31 <@zere> uhhh... i assume trac can search by keyword...
18:31 < TomH> hmmm actually that seems to require configuring the keywords in the config file which isn;t very good
18:31 <@zere> but i'm looking at the search page and it's not clear quite how.
18:31 < TomH> I assumed it would offer what was already uises
18:32 < SteveSn_> http://trac.openstreetmap.org/query?status=accepted&status=assigned&status=new&status=reopened&component=website&order=priority&col=id&col=summary&col=status&col=owner&col=type&col=priority&col=milestone&col=changetime&keywords=~foo
18:32 < SteveSn_> will give you all the tickets with the keyword foo
18:33 <@zere> ok. so we can have a wiki page linking to a list of open tickets with a particular keyword, but it seems we need to know the keywords in advance.
18:33 < emacsen> is there a way to ask trac for a list of keywords. Tom mentioned a cloud?
18:33 < TomH> did I ?
18:34 <@zere> i suggest we all spend some time going through trac and next week we'll have a better idea of what keywords we'll need. can i get some +/-1s?
18:34 < emacsen> +1
18:35 < SteveSn_> +1
18:36 < emacsen> TomH, sorry, I thought I read that. But I'll put the crack pipe down now
18:36 < apmon> +1
18:36 <@zere> awesome. :-)
18:37 <@zere> so other suggestions for stuff we could do. let's quickly go through some of the options.
18:37 <@zere> +1/-1 on whether developer-oriented videos would be helpful?
18:37 < TomH> eerggh
18:38 <@zere> s/eerggh/-1/?
18:38 < apmon> would any?
18:38 < TomH> -1000
18:38 < TomH> but hey maybe that's just me
18:38 < emacsen> zere, can you elaborate
18:38 < emacsen> what would a developer oriented video be of?
18:38 < SteveSn_> I doubt I'd watch a developer video or at least not until I've exhusted text docs. -1?
18:38 < apmon> I'd usually rather go through written instructions at my own speed, than to listen to tutorial videos
18:39 < TomH> I know my reaction to finding that, when I investiagte some software or technology the main source of information is a video is to close the web site and go find something more useful to do
18:39 < TomH> video's are just too low bandwidth
18:39 <@zere> installation steps of the rails_port, perhaps. making a change to the style/CSS, etc... i dunno, just casting around here. 18:39 < emacsen> my tutorial videos have 200 views in 2 weeks
18:39 < TomH> I can get an overview about 100 times quicker from text than from video
18:39 <@zere> that seems like a pretty uniform -1, then ;-)
18:39 < TomH> but hey, they may be people it works for
18:40 < TomH> in fact given how bloody common it is these days I really hope there is somebody that it works for ;-)
18:40 < emacsen> I think simple screencasts can be helpful, but that's me
18:40 <@zere> something interactive perhaps? +1/-1 on IRC workshops? i mean that's kinda the dev channel, but i mean setting a time dedicated for people to help newcomers.
18:41 < emacsen> http://www.youtube.com/playlist?list=PLCE6296A33CF47955&feature=mh_lolz shows the first video has >250 views, second one 100. That's in about 2 weeks, so I'd say that the format is mildly popular at least
18:41 <@zere> emacsen: yeah. not much return on investment :-(
18:41 < TomH> teaching people to edit the map is a very different use case though
18:41 < emacsen> -1 They take up a ton of time, people can't make it, then they complain, they go slow etc.
18:42 < TomH> and I'm not sure the viewer counts mean much as most new users will have no idea they exist
18:42 <@zere> yeah. seems like making the wiki documentation better is priority number 1.
18:42 < emacsen> I'm already going through the rails installation
18:43 < SteveSn_> -1, to the IRC workshop, because I don't think helping 5 people at once will work well. helping them individually on IRC might work better
18:43 < emacsen> as was someone else on IRC, but I forget who
18:44 <@zere> ok. another unanimous -1. anyone have any idea they'd like to discuss?
18:45 < apmon> Is there going to be any developer related activity at SotM?
18:45 < apmon> I.e trying to recruit developers to some of the core software infrastructure?
18:46 <@zere> i don't see anything on the plan other than the tech session and some assorted talks, but we could try and shoehorn something in.
18:46 < SteveSn_> a lightning talk aimed at developers on some components of the software?
18:46 < apmon> Or even explaining to the general audience what all the components involved are?
18:46 <@zere> a lightning talk could work. who's going to be at sotm? i am.
18:47 < TomH> not me
18:47 <@zere> "developing OSM 101"
18:47 < SteveSn_> I will be there (but I don't know enough about the rails_port to give a talk on it)
18:47 < apmon> I am
18:49 <@zere> there's an "unconference" session. dunno what's going to fill that space.
18:49 < apmon> zere: I guess we could put something together for a lightning talk or two
18:49 <@zere> apmon: ok. we should do that :-)
18:51 <@zere> any other ideas? open season, come on guys!
18:52 < apmon> ideas in general or related to sotm?
18:52 < emacsen> I don't think we should bite off more than we can chew
18:52 < apmon> ideas in general: An irc based thematic hack-weekend
18:53 < apmon> It isn't really any different than normal osm-dev, but by declaring it a "hack-weekend" and announcing it, it might increase motivation and participation
18:53 <@zere> anything! sure, we shouldn't bite off too much, but i don't want to miss out on some good ideas because we were too focussed on something else.
18:54 <@zere> apmon: to coincide with a real-life hack weekend, or a pure e-hack-weekend?
18:55 < emacsen> I thought when there's a live hack weekend, everyone's on IRC anyway
18:55 < apmon> possibly pure e-hack. Otherwise people in a RL meeting aren't inclined to communicate via irc
18:55 <@zere> i dunno... last hack weekend there was a lot of screen-staring ;-)
18:56 <@zere> apmon: are you volunteering to organise it?
18:56 < apmon> Well, it is something we can try out either way and see how it goes
18:57 <@zere> usually i tempt people with free pizza - i'm not sure what the IRC version of that is ;-)
18:57 < apmon> zere: Not having to pay for travel... ;-)
18:57 < apmon> I can try and organise a e-hack weekend.
18:58 <@zere> apmon: great, thanks :-)
18:58 < apmon> The biggest issue is probably to find a common theme that enough people are interested in to make it a project
18:58 < apmon> I think the i18n rails_port hack-weekend was a great example of it
18:58 < apmon> where quite a number of additional people joined and helped over IRC
18:58 <@zere> anyone else want to discuss something? mkl1, RichardF, you've been very quiet.
18:58 < SteveSn_> let us see what themes come out of classifying bugs into keywords
18:59 < mkl1> hey been half watching, half on the phone
18:59 <@zere> yes, i think that could be a useful thing. otherwise taking on 1,100 issues on trac really does seem like trying to climb everest in a weekend.
18:59 < mkl1> only thing i have to say, i was going to email the design@ list to get some conversation started on user requirements, but haven't done it yet
19:00 < emacsen> user requirements?
19:00 < mkl1> and it's been fun to watch my 6 year old trac issues get closed ... thanks
19:00 < mkl1> emacsen: i just mean, thoughts on features and designs for the site. what kinds of things would serve users
19:01 <@zere> mkl1: do you think some were closed that shouldn't have been?
19:01 < emacsen> so, I see why that's valuable to the design folks, but is that valuable to this group?
19:01 < mkl1> zere: looked good to me
19:01 <@zere> emacsen: because there's going to be overlap with what we're trying to do.
19:01 < emacsen> ah, gotcha
19:02 < mkl1> we talked about this last week, about how to engage in a useful user conversation
19:02 <@zere> and (imho) many of the guys who'll have useful design input aren't going to be postgres/rails experts.
19:02 < RichardF> zere @ 18.58: apologies, quiet due to train being cancelled and being delayed back mutter mutter
19:02 < mkl1> i think it's worth a shot to engage this list ... includes folks like Stamen, who focus a bit more on the user experience, but are also super informed
19:03 <@zere> but i want them to be able to contribute as well as any other beginning developer.
19:03 < emacsen> mkl1, I'm just not sure if we're ready for that conversatio yet. We need to clean house
19:03 < mkl1> right, we can facilitate collaboration
19:03 <@zere> emacsen: don't worry, i don't think anyone wants to rush anything. there's time to get it right.
19:04 < mkl1> emacsen: if we possibly got to the point where that would be a hassle for you, i consider it a win
19:04 <@zere> mkl1: has there been any discussion of Firefishy's "design tournement" idea?
19:04 < emacsen> mkl1, what do you mean?
19:04 < emacsen> "a hassle for me"?
19:04 < mkl1> not because i want to hassle you, but because it means that things are moving? don't worry...
19:04 <@zere> i think he means that it's a good problem to have that level of activity.
19:04 < mkl1> right
19:04 < emacsen> I just don't want to invite new developers to a mess.
19:05 < mkl1> zere: it was supposed to be discussed at the management meeting last week, but i missed it
19:05 < emacsen> once we can give them a positive experience, that'l be better to start the conversation
19:05 <@zere> yeah, me too. maybe we can have one pre-sotm? are most of us going to be there in-person?
19:05 < mkl1> emacsen: i think this will be understood. it's more to just get some dialogue there, not pelt trac with a bunch of requests
19:05 <@zere> emacsen: i agree.
19:06 < emacsen> I am unlikely to be at sotm. My schedule's a mess and I'd need to come back to town and leave the next day, and I haven't bought anything yet
19:06 < TomH> zere: I suspect not given that relatively few UK people are going
19:06 < mkl1> i'll be in denver from 7-15
19:06 < mkl1> but i'm merkan
19:06 <@zere> and given my experience setting up rails_port, i think that's something that we'll have to discuss with anyone interested in doing CSS-type work.
19:07 <@zere> i'll be there and (iirc) harry will be there. anyway, might be good to have a "meeting" even if it's not official.
19:08 < TomH> yes I think it's just you and harry from the London mob
19:08 < emacsen> zere, yeah so I hope to have something practical next week
19:08 <@zere> (^^ given that i suspect many people who do design prefer not to work on linux :-( )
19:08 < emacsen> though... shit. I won't be able to make next weeks's meeting
19:09 < emacsen> I just realized it's the 5th. I have a meeting that time
19:09 <@zere> i think i should have sent a reminder out to dev@ again... attendance is way down...
19:11 < apmon> zere: First time is always more busy than the following meetings