Humanitarian OSM Team/Working groups/Technical/meeting 2014-05-05

From OpenStreetMap Wiki
Jump to: navigation, search

Meeting to discuss Tech topics on Monday 5th May 2014

IRC log:

19:00:32< dodobas> hmm, it looks like it's going to be an intersting meeting
19:01:17< dodobas> first of all, i have to apologize... for not doing the TLDR summary for the last meeting
19:02:05< dodobas> anyway... does anyone wants to start the meeting
19:02:29-!- larryone [~larryone@] has quit [Quit: This computer has gone to sleep]
19:02:33< AndrewBuck> IS there a hackpad for an agenda or anything?
19:02:39< dodobas> usual format is that you apply for the 'mic' and then you report/raise issus
19:02:42< mkl1> hey dodobas: i'm here, but thought we could get into the OSMTM discussion after the rest of the agenda is gone through
* clara is reverting features so I give you my update when the link works 19:02
19:03:09< mkl1> i'm not sure if pierregiraud, and MoniqueMitchell, and others here for OSMTM have all gathered yet
19:03:10< dodobas> AndrewBuck: we don't have a formal agenda, maybe we should start preparing
19:03:44< mkl1> pierregiraud and i started preparing this for the osmtm topics ...
19:04:08< dodobas> mkl1: k, we can hold OSMTM then...
19:04:23< dodobas> maybe I could start with my service report...
* pierregiraud is here 19:04
19:04:39< dodobas> and then we could move onto OSMTM and other reports
19:04:47< dodobas> so...
19:05:07< mkl1> sounds good to me dodobas
* MoniqueMitchell is here 19:05
19:05:52< dodobas> in the last month I have updated OpenAtrium2 setup and enabled 'reply by email' module/feature
19:06:02< dodobas> OA2 is used by the HOT-board
19:06:56< dodobas> I've also reconfigured OSMTMv1 service to auto-kill/restart zombie processes... making it more durable
19:07:31< dodobas> as there are some issues when admin users try to create new jobs using osm relation ids...
19:08:24< dodobas> and those are hard to debug, as sometimes the OSM data download service refuses to return valid data
19:08:27-!- Bilderhobbit [] has joined #hot
19:08:49< dodobas> I've also added South-america dataset to the export server
19:09:01< dodobas> *hot export service
19:09:25< dodobas> as there was a requrement for data exports
19:09:41< AndrewBuck> so the export server is not worldwide then?
19:10:24< dodobas> I've also tried to create a testing instance of the export server, so designers could implement new export server look (there is an issue on github for it)
19:10:42-!- JulienME [~Julien@2a01:e35:8b60:e720:21f:d0ff:fe27:7e0e] has joined #hot
19:11:23< dodobas> but there are several problems, one of which is lack of my ruby/rails application development/deployment experience
19:11:41< dodobas> but I should be able to do it, eventually
19:12:01-!- larryone [~larryone@] has joined #hot
19:12:07< dodobas> there was also a problem with disk space usage on export server
19:12:25-!- larryone [~larryone@] has quit []
19:12:37< dodobas> so I've implemented a data purging policy... (let me find the issue)
19:12:55< mkl1> dodobas: a few questions/feedback when you're ready
19:12:56< dodobas>
19:13:08< dodobas> mkl1: just one more thing...
19:14:09< dodobas> I've also FINALLY switched export server to the public Github repo... we were still using private repo of the company which initially developed the service
19:14:41< dodobas> and it was almost impossible to merge changes from the public repo...
19:14:48< dodobas> and FINAL thing...
19:14:49< AndrewBuck> The policy on that issue 70 looks good to me.
19:15:21< dodobas> we got HOTOSM Transifex team, and i've initially created hot-export translation project
19:16:17< dodobas> the nices thing is that we can almost automatically manage translation using the command line client
19:16:35< dodobas> and... im done... any comments, mkl1, AndrewBuck
19:17:02< pierregiraud> we’ll probably talk about transifex later for the tasking manager
19:17:09< dodobas>
19:17:16< AndrewBuck> Yeah, so that policy for deletions looks pretty good, one thing I would add is that if you lok at existing jobs the spatiallite and sqllite files are way bigger than the others...
19:17:34< AndrewBuck> if we avoid creating these in the first place we wouldn't be out of diskspace.
19:17:55< mkl1> that's great to see action on the export server. are there any other devs out there interested to help, get you up to speed on deployment, and to implement things like the design improvements?
19:17:58< AndrewBuck> Unchecking them by default in the job creation page would be a good idea I think/
19:18:11< dodobas> AndrewBuck: i've bene told that export service was never designed to be 'history data download sevice'
19:18:34< mkl1> wondering how to advertise and mobilize this (and other top HOT tech needs)
19:19:05< dodobas> mkl1: i got some pointers... so I should be good to go (for depolyment and fixes)
19:19:08< AndrewBuck> dodobas: yeah, And that is why the deletion policy makes sense, I am just saying that on top of that almost no one uses the spatialite or sqllite products so making them by default just makes jobs slower to run and also way bigger on disk
19:19:23< AndrewBuck> The should be opt-in rather than opt-out
19:19:32< mkl1> also, do you have a sense of system resources it would take to expand geographic areas further? not saying we need to, just would be a good idea to know what it might take if so
19:20:04< mkl1> actually, someone already making pull requests :)
19:20:36< dodobas> mkl1: first of all it will add more overhead on the constantly updating osm data changes
19:20:41< mkl1> final question, just curious, what do you use to monitor and restart zombie processes? just a cron script or ... ?
19:20:58< dodobas> currently it takes 15-20 minutes for one run
19:21:29< dodobas> it also takes longer to extract data from a larger file...
19:21:37< mkl1> 15-20 to digest how long a time period ... one hours worth of changes?
19:21:46< dodobas> last 15-20 min
19:22:06< AndrewBuck> so it is barely keeping up then?
19:22:09< mkl1> so it's only just keeping up?
19:22:18< dodobas> so... expansion is not a problem if we keep the areas small
19:22:25< dodobas> *supported areas
19:22:40< dodobas> well, the data is 'stale' for max 20min
19:22:57< AndrewBuck> oh, I see what you are saying.
19:23:00-!- larryone [~larryone@] has joined #hot
19:23:15< AndrewBuck> so it updates minutely but sometimes is as much as 20 minutes behind the main OSM server?
19:23:16< dodobas> mkl1: for osmtm... we use 'uwsgi' app server which can check for process state and 'restart' it
19:23:22< dodobas> AndrewBuck: yes
19:23:47< dodobas> it was designed to acocmadate fast data updates for small areas
19:23:51< AndrewBuck> ok, would be nice to show the timestamp of the latest file on the page so we know how up to dat it is.
19:24:08-!- larryone [~larryone@] has quit []
19:24:57< dodobas> AndrewBuck: well if you create an issue, it might get done... something like 'platform status info box'
19:25:53< dodobas> IMHO, export server is functional but there are a lot of issues
19:25:53-!- robertsoden [] has joined #hot
19:26:06< dodobas> some of which are design issues
19:26:21< mkl1> would it be something to consider for raising funds and hiring a dev to focus on for a while?
19:26:22< pierregiraud> UI or backend?
19:26:36< dodobas> for example you can create a 'new job run' or 'create a similar job'
19:26:37< mkl1> or do you think something that we can organize contributions to get in good shape?
19:27:00< dodobas> in the end we have a lot of 'one run simliar jobs' created just a few minutes apart
19:27:27< dodobas> mkl1,pierregiraud maybe we should focus on redesign/design for now
19:27:49< dodobas> to make it more clear what a button does, and how it should be used
19:28:16< mkl1> the redesign from sotmus sprint looked pretty good
19:28:54< dodobas> i've noticed that there is some functionaliy that was planned but never completed
19:29:25< dodobas> or maybe it works, but it's so hairy that noone knows how to use it properly
19:29:45< pierregiraud> are guys from geofabrik still involved in the project?
19:29:52< dodobas> like... database translation files... which should translate tags between projects
19:30:03< AndrewBuck> yeah, the tag translation stuff is a complete mystery to me. I think it actually does somethng useful but no one has really set it up.
19:30:33< AndrewBuck> That is one of the things the ontology I am working on is supposed to help with.
19:30:43< dodobas> pierregiraud: not that i know of... there were no significant commits after initail service deployment
19:30:57< dodobas> some fixes... but it's hard to track their private repo
19:30:59< mkl1> yea, and it shouldn't be exposed to most users. right now, it's just a distraction in the UI for most folks
19:31:15< dodobas> i could research a bit... if you are interested
19:31:38< mkl1> so is hot-exports an orphaned project? or dodobas, have you taken "ownership" of managing what happens with it?
19:31:50< AndrewBuck> dodobas: research? was that to me or pierregiraud?
19:32:09< dodobas> AndrewBuck: it was fr pierregiraud
19:32:13< AndrewBuck> ok
19:32:24< AndrewBuck> thats what I thought but wasn't sure
19:32:50< dodobas> mkl1: I can maintain it... merging testing(maybe)... but I'm not a ruby developer...
19:33:03< mkl1> ok
19:33:36< dodobas> some crazy BIRD gave me admin access on HOT github repo... so i could be an 'owner' .. :)
19:34:44< dodobas> i could 'play' with development, but only in the context of small issues/features
19:35:07< mkl1> (btw, i probaby just have 45 minutes more here, and would like to start on osmtm stuff. oh and also need to raise request for write up on our working group mandate, etc)
* dodobas should at least finish Rails tutorial 19:35
19:35:31< dodobas> mkl1: please...
19:35:31< clara> Just a very short update from me: I've worked on proper multilingual version of - it needs manually re-adding the translation for 8 updates, and then translating the menu items etc to which ever languages ppl want
19:35:38< clara> ... end
19:36:25< dodobas> clara: is that facilitated on the site... or you need to use some external l10n service ?
19:36:30< dodobas> like transifex?
19:36:35< clara> it's all on the site
19:36:38< dodobas> k
19:36:42< mkl1> (i'll raise the mandate question next. and then if go-rounds have finished, osmtm)
19:36:56< dodobas> mkl1: you have the hammer
19:37:24< mkl1> yea, so a bit of housekeeping / org dev stuff
19:37:45< mkl1> we've started formalizing the working groups structures. you all may have seen some emails about that on hot list
19:38:01< mkl1> tech working group is the only continually functioning one, so congrats to us :)
19:38:14< mkl1> anyway, board is asking for a few simple things
19:38:16< pierregiraud> thanks to dodobas
19:38:25< mkl1> totally :)
19:38:58< mkl1> 1) "mandate" of the group. basically what is the purpose of the group, what is it responsible for, what does it hope to accomplish
19:39:03< mkl1> in broad strokes
19:39:42< mkl1> 2) how is the group organized? ie meeting frequency, where, who leads the group, where are logs and updates kept, what other tools used to organize (ie github)
19:39:58< mkl1> that's pretty much it. written up in the wiki i guess
19:40:13< mkl1> i think we have a clear idea of this, just needs to be all explicit
19:40:25< mkl1> anyone want to take a turn at doing this write up in the wiki?
19:40:32< dodobas> bullet points FTW :)
19:40:47< mkl1> i'm responsible to share back with the board at next meeting in 2 weeks
19:41:08< dodobas> we could start a hackpad document
19:41:18< dodobas> it may be eaiser to collaborate
19:41:20< mkl1> sounds good dodobas
19:41:33< mkl1> we can work it out there
19:41:54< mkl1> i'll leave it at that. we can pick this up outside of meeting...
19:41:59< dodobas> k
19:42:13< mkl1> is there anything else on agenda or are we ready to jump into osmtm?
19:42:59< ptressel> I'm waiting til after to talk to AndrewBuck re JOSM...
19:43:37< dodobas> mkl1: let's jump into osmtm
19:43:39< mkl1> k
19:43:57< mkl1> so pierregiraud organized topics into this hackpad
19:44:03< mkl1>
19:44:09< mkl1> from a high level, there's three main topics i think
19:44:17< mkl1> 1) getting osmtm v2 to launch
19:44:43< mkl1> 2) coordinating a software management process from here on
19:45:01< mkl1> 3) starting a design and ux discussion
19:45:20< mkl1> pierregiraud, do you want to take the "hammer" from here?
19:45:34< pierregiraud> ok, but don’t hesitate to help me when required
19:45:53< pierregiraud> first of all, as written by Mikel in a mail to hot list
19:45:54< mkl1> for sure, i'm here (with a baby actually ...)
19:46:14< pierregiraud> I had time to work on the V2
19:46:32< pierregiraud> there was a list of features from V1 to put in V2
19:46:49< pierregiraud> this first phase looks complete
19:47:02< pierregiraud> everything is in now
19:47:26< pierregiraud> I created a new tag 2.1
19:47:39< pierregiraud> of course there are probably some bugs left
19:47:56< sev_hotosm> pierregiraud I still owe you a buffer layer for testing the polygon import
19:48:05< pierregiraud> I paid attention to test coverage
19:48:12< pierregiraud> you’re right sev_hotosm
19:48:19< pierregiraud> but I’m confident
19:48:35< pierregiraud> I also added a file for the License
19:48:46< pierregiraud> which I chose to be 2-Clause BSD
19:49:02< pierregiraud> I hope this suits HOT needs
19:49:24< pierregiraud> I’m now looking for testers
19:49:24< dodobas> anything permissive is fine by me ... BSD/MIT/...
19:49:31< pierregiraud> both as users and admins
19:49:41< mkl1> i think there were no objections to bsd license
19:49:59< sev_hotosm> pierregiraud I am in for testing
19:50:10< mkl1> here's the current staging instance
19:50:15< clara> I can test as well
19:50:23< pierregiraud> great
19:50:38< mkl1> we were talking about getting it set up on a hot server at this stage of testing
19:50:49< pierregiraud> the demo instance was kept secret but now I think the more users we have to test the better
19:51:02< pierregiraud> mkl1: correct
19:51:25< pierregiraud> this would also be a good way to know if installation procedure works
19:52:14< mkl1> i think dodobas said he was ready to do this
19:52:14< pierregiraud> once again testers ping me if you need admin privileges
19:52:23< sev_hotosm> yes please
19:52:24< dodobas> yeah...
19:52:49< dodobas> do we want to move repo to the hot GH organization, or not yet ?
19:52:53< pierregiraud> dodobas: I hope the procedure is better documented than it was for V1
19:52:56< mkl1> pierregiraud: how do you want to manage user bug reports and feedback? some will be 2.1 appropriate, others longer term
19:53:04< clara> pierregiraud: if there are specific issues you want to have tested, maybe you could make a list?
19:53:10< sev_hotosm> so as you told me in SOTMFR, no more html? :)
19:53:38< pierregiraud> mkl1: there are milestones in github
19:54:00< dodobas> mkl1: do we want to take that direction or just call everything 'current'
19:54:01< pierregiraud> I think it’s repo manager(s) who should label issues
19:54:10< pierregiraud> and add them to milestones
19:54:18< dodobas> its not like we are going to run different version s
19:54:31< mkl1> who is the repo manager
19:54:35< pierregiraud> sev_hotosm: HTML is still possible but not recommended
19:54:51< pierregiraud> mkl1: currently I am
19:55:00< pierregiraud> but others can help too
19:55:25< pierregiraud> this looks like we are already in the « coordinating project » of the agenda
19:55:29< mkl1> i'm willing to keep helping. if we can organize regular times to check in and coordinate
19:56:21< pierregiraud> clara nothing to be tested specifically
19:56:26< clara> k
19:56:27< dodobas> so, my two cents....
19:56:39< pierregiraud> dodobas: I’m OK to move to HOT github org
19:57:02< dodobas> milestones on GH are only useful if you have a company 'mindset'
19:57:27< dodobas> i.e. you can set a release date... and work towards that data
19:57:42< mkl1> it's appropriate now, since camp2camp is engaged on this. but not necessarily after that
19:57:59< pierregiraud> I understand
19:58:06< dodobas> we should probably 'design' coordination scheme to be more organic
19:58:25< pierregiraud> are milestones == tags/versions in your opinion?
19:58:45< dodobas> tags are useful for big code pushes
19:58:57< pierregiraud> I agree
19:59:04< pierregiraud> and for data migration
19:59:20< pierregiraud> I like the idea of maintaining a changelog file too
19:59:36< dodobas> constatnly evolving project, based on community efforts, ..., well it's hard to tell
20:00:29< dodobas> we could, also as part of the 'madate', create bi-monthly project status review
20:00:44< dodobas> which would be a responsobility for project managers
20:00:44< pierregiraud> looks good to me
20:00:59< mkl1> that makes sense to me
20:01:32< dodobas> and then publish a changelog file, and tag a repo if its reasonable
20:02:24-!- Robert_Banick [] has joined #hot
20:02:30< mkl1> +1
20:02:33< dodobas> just a simple report... 10 issues are opened, 2 have been closed, high priorities are...
20:02:52< pierregiraud> +1
20:03:01< dodobas> which could be intergated with a general 'call for help' HOT documents
20:03:21< dodobas> to attract new develpers
20:03:30-!- cquest_ [] has quit [Remote host closed the connection]
20:03:33< sev_hotosm> how do we hierarchize priorities
20:03:44< sev_hotosm> regarding feature requests?
20:03:52-!- cquest_ [] has joined #hot
20:03:58< dodobas> it would be reposnibility of the project manager
20:04:12< pierregiraud> is there only one project manager?
20:04:15< dodobas> disscused on TWG meetings, or something similar
20:04:45< mkl1> i think there could be more than one, but not too many
20:04:47< pierregiraud> or during those bi-monthly project statuses
20:04:48< sev_hotosm> because it is not something purely technical
20:05:06< dodobas> pierregiraud, mkl1 more than one per project ?
20:05:30< pierregiraud> yep
* pierregiraud just asking 20:05
20:05:54< pierregiraud> it’s OK if there’s only one project manager
20:06:17< dodobas> pierregiraud: as long as the report gets done... i don't really care
20:06:48< mkl1> sev_hotosm: it would still be up to a project manager to organize requests accordingly
20:07:06< dodobas> sev_hotosm: well, that could also be part of the report... there are 5 outstanding new feature request... we should coordinate with the funding wg and see if we can make it happen
20:07:31< sev_hotosm> maybe the other Working Groups could internally discuss and decide what are their main needs regarding OSMTM improvements (if any) and hierarchize them. So the dev team would have the requests of every WG ?
20:07:56< pierregiraud> looks good too
20:07:58< mkl1> and it really depends on the situation, re: is it organized under a funded project, with specific requirements? or top HOT needs could be highlighted, for developers to contribute...
20:08:13< sev_hotosm> and then the project manager would prioritize regarding feasibility/emergency/funds, etc. ?
20:08:20< dodobas> sev_hotosm: project manager allows you to do that, as you have a 'person' to communicate
20:08:36< mkl1> the only appropriate working group might be activation working group?
20:08:43< sev_hotosm> not only
20:08:59< mkl1> funding would be relevant too, but for not necessarily for setting priorities
20:09:24< sev_hotosm> TWG also eg to imrpove technical part
20:09:42< sev_hotosm> AWG yes will be maybe the main user
20:09:43< pierregiraud> I think we all agree that the idea is to do our best to get OSMTM improved
20:10:02< mkl1> sure TWG would be home to software project management, makes sense
20:10:04< sev_hotosm> but Communciation WG might also have specific requests
20:10:45< mkl1> ok
20:10:51< dodobas> sev_hotosm: that should be discussed on the board meeting, right ?
20:11:18< sev_hotosm> dodobas I think it is a community related topic
20:11:41< dodobas> once the board has all the mandates, and can cross check and correlate where do WG interact and overlap?
20:11:56< sev_hotosm> how the TWG interacts with the other WG
20:12:06< mkl1> i think all sev_hotosm is saying that if a working group has some request regarding the osmtm specifically
20:12:28< mkl1> then that would go through the osmtm project manager(s)
20:12:49< mkl1> who would be coordinated such things within the twg checkins
20:13:21< skorasaurus2> back :)
20:13:26< pierregiraud> can we move on?
20:13:37< mkl1> ok by me
20:13:49< pierregiraud> what about the github repo? is it OK to move to HOT org?
20:13:51< sev_hotosm> yes
20:14:11< sev_hotosm> I meant yes for we can move on
20:14:12< pierregiraud> it’s currently in my public repo
20:14:39< pierregiraud> in my opinion, we can move it to HOT org already
20:15:11< mkl1> (though 1 thing to note for future discussion , is how to handle requests in the middle of a crisis situation. would need to ensure adequate resources to manage emergency requests, which can move fast)
20:15:38-!- cquest_ [] has quit [Remote host closed the connection]
20:15:38-!- cquest_ [~cquest@2a01:e35:2e69:4cf0:a441:83d3:f847:d22d] has joined #hot
20:16:06< pierregiraud> dodobas: is it difficult to change the repo location?
20:16:15< pierregiraud> do I need to do something specific?
20:16:24< dodobas> pierregiraud: i'll need to check...
20:16:43< dodobas> there is probably somthing in the repo settings, ill get back to you
20:16:51< pierregiraud> OK
20:17:00< pierregiraud> next point is translations
20:17:15< pierregiraud> dodobas: are you confident with transifex?
20:17:20< dodobas> yes
20:17:30< sev_hotosm> translations is for the UI or the jobs?
20:17:31< pierregiraud> OSMTM V2 is using gettext .po .mo files
20:17:41< pierregiraud> translations for the UI
20:18:42< pierregiraud> is transifex configuration easy to set up?
20:19:05< dodobas> pierregiraud: i can set it up, and add documentation for it
20:19:27< pierregiraud> what about as proposed on the mailing list?
20:20:04< pierregiraud> I’ve used transifex for the Kort project and I really liked it
20:20:08-!- alexbarth [] has quit [Remote host closed the connection]
20:20:24< dodobas> do they have a CLI utility which allows you to sync new transltions with a single command :)
20:20:41-!- alexbarth [] has joined #hot
20:20:54< pierregiraud> I don’t know but I would be surprised it’s the case
20:21:52< dodobas> so, translate wiki might have a larger community of translators...
20:22:31< dodobas> but I would still recommend transifex... it's even opensource
20:22:32-!- alexbarth [] has quit [Read error: Connection reset by peer]
20:22:33< pierregiraud> the problem is that we want people who know what the application talks about translate it
20:22:44< pierregiraud> +1 for transifex
20:22:55-!- alexbarth [] has joined #hot
20:23:22< pierregiraud> about statistics now
20:23:35< mkl1> pierregiraud, one thing
20:23:53< pierregiraud> Kate created an issue on v1 to switch from google analytics to piwik
20:24:09< pierregiraud> mkl1: yep?
20:24:14< mkl1> i have to run in a second .. just wanted to say that i think this one is not necessarily 2.2, but might be a 2.1 ... ie duplication of existing v1 feature
20:24:53< mkl1> (also, I know MoniqueMitchell and robertsoden have been patiently waiting to talk about the design initiative at CU Boulder)
20:24:56< mkl1> that's all...
20:25:04< mkl1> sorry to interrupt
20:25:05< pierregiraud> let’s talk about this one later
20:25:11< pierregiraud> I’ll ping you tomorrow
20:25:37< pierregiraud> thanks mkl1
20:26:10< dodobas> pierregiraud: what are your thoughts on GA vs. piwik ?
20:26:26-!- sev_hotosm2 [] has joined #hot
20:26:30< pierregiraud> for the analytics? can we still keep things logged in GA and get piwik set up later?
20:27:06< pierregiraud> I know that people would like us to use google as less as possible
20:27:14-!- robertsoden [] has left #hot []
20:27:16< clara> run it paralell for a bit to get an idea of the differences
20:27:24< pierregiraud> I kinda like GA
20:27:29< pierregiraud> clara: good point
20:27:33< dodobas> pierregiraud: i don't have any preferences...
20:27:38-!- robertsoden [] has joined #hot
* clara regulary uses piwik for sites and I'm happy to show you what's possibly with i 20:28
20:28:09< pierregiraud> I don’t know piwik much, but I like the real-time tab in GA
20:28:17< dodobas> we can run thosein parallel
20:28:23< pierregiraud> ok great
20:28:45< pierregiraud> I think we’re done with the 2 point in mkl1’s agenda
20:28:49< dodobas> users' browser will just create another request...
20:29:52< dodobas> pierregiraud: at which point should I do a code review?
20:30:03< pierregiraud> good question
20:30:19< pierregiraud> I think the code is already ready for a review
20:30:32< dodobas> initially it should have happend after the code 'delivery'
20:30:36< pierregiraud> how do you want to proceed?
20:30:50< pierregiraud> the code is not really delivered yet
20:30:57< pierregiraud> and there’s a 2nd phase
20:31:04< pierregiraud> upcoming
20:31:30< mkl1> i think we can aim to deploy following completion of 1st phase
20:31:30< dodobas> hmm, then maybe before the 2nd phase starts
20:31:46< pierregiraud> sounds good
20:31:55< mkl1> dodobas, is 1st phase deployment dependent on code review or not?
20:32:17< dodobas> mkl1: it's not, CR is just a bonus :)
20:32:23< mkl1> ok
20:32:45< pierregiraud> dodobas: code is pep8 and flake compliant now :-)
20:32:50< pierregiraud> :-p
20:32:59< mkl1> what would be the timeline for CR? we're discussing phase 2 timeline with camp2camp tomorrow
20:33:20< dodobas> i probably need couple of days...
20:33:28< mkl1> ok
20:34:24< pierregiraud> OK, time for the last point
20:34:35< dodobas> again ... is not something that should hinder any new development
20:34:51< pierregiraud> « design project » for University of Colorado students
20:35:04< pierregiraud> MoniqueMitchell: you around?
20:35:16< MoniqueMitchell> yes
20:35:40< pierregiraud> do you want to introduce yourself and talk a bit about this nice project?
20:36:27< MoniqueMitchell> Howdy everyone
20:36:58< MoniqueMitchell> My name is Monique & I'm a graduate student taking a Social Computing class at CU.
20:37:05< dodobas> TX :)
20:37:51< skorasaurus2> howdy.
20:38:50< MoniqueMitchell> A group of myself and 5 other graduate students have been looking at the TM /HOT org with the hopes of making some recommendations with the hope of improving the user experience..
* clara waves goodbye 20:39
20:39:35-!- clara is now known as clara-afk
20:40:00< MoniqueMitchell> We've been focusing primarily on organization and social aspect and how those can make onboarding and validation less painful processes.
20:40:44< MoniqueMitchell> Our team is made up of a group with a diverse background--from computer scientists, to designers, to linguists, to photo journalists.
20:42:00< MoniqueMitchell> Our project is a preliminary review and will actually be wrapping up in the final day or two, at which time I'll be sending over the final documentation complete with a medium fidelity mockup of the changes we're recommending to the listserv.
20:42:37< dodobas> MoniqueMitchell: whats 'a medium fidelity mockup' ? :)
20:42:55< MoniqueMitchell> For now, we have some initial ideas here if you are interested:
20:42:59< pierregiraud> I’ve seen the draft for this document and there’s already a lot of interesting things
20:43:22< MoniqueMitchell> We're working on getting some wire-frames that are hosted on a groupmate's server
20:43:28< pierregiraud> though some may not be realisticly added to the tasking manager easily
20:43:37< MoniqueMitchell> basically, a website that is imperfect code to visualize some of our ideas
20:44:38< pierregiraud> MoniqueMitchell: how can we help you before you write the final document?
20:45:14< MoniqueMitchell> We had two types of ideas: wild crazy ideas that may not be integrateable and smaller simpler ideas that could be implemented. We decided with this project to take on even the wild and crazy ideas since coming from an outside perspective one can often see things that may not be appearant.
20:46:18< MoniqueMitchell> pierregiraud: I've received some comments already from the listserv directed to myself personally, but if anyone wants to take a look at the preliminary document and answer any questions/pose their own questions, that would be helpful.
20:47:09< pierregiraud> one stupid question: how did you choose the tasking manager and HOT to be the target for your project?
20:47:22< MoniqueMitchell> Because of the nature of a semester project, huge design changes likely cannot be incorporated at the moment, but knowing where the questions are will help us better write-up our ideas so that the TMWG/HOT understands what we are trying to say.
20:48:03< MoniqueMitchell> The EPIC project at CU, which our professor runs, works with Twitter data in disaster situations.
20:48:24< MoniqueMitchell> As such, our class has had a heavy humanitarian computing aspect.
20:48:29< pierregiraud> is there a phase in your project when you’ll actually dig into the code and develop things?
20:48:53-!- Bilderhobbit [] has quit [Quit: CGI:IRC]
20:49:47< MoniqueMitchell> This project was structured purely as a design project. However, I specifically asked to be placed on this team because I would be interested in continuing to work on the TM in a more technical function once the semester is over.
20:50:23< pierregiraud> great
20:50:36-!- cquest__ [] has joined #hot
20:51:02< MoniqueMitchell> I know that with TMv2 the idea was to make the code easier for other developers to join in with. I'd be interested in implementing an idea or two that came out of this project, as would my younger sister Sarai Mitchell.
20:51:08-!- sev_hotosm2 [] has quit [Ping timeout: 480 seconds]
20:51:28< pierregiraud> contributors are welcome
20:51:38< MoniqueMitchell> She couldn't be on this chat as she is travelling at the moment.
20:51:51-!- cquest_ [~cquest@2a01:e35:2e69:4cf0:a441:83d3:f847:d22d] has quit [Read error: Connection reset by peer]
20:52:18< MoniqueMitchell> Anyhow, that's the quick summary of what we've got going on. Any other question/things y'all would like me to flesh out?
20:52:38-!- jgc [] has quit [Quit: Quitte]
20:52:41< pierregiraud> I’m looking forward to the final result and the wireframes
20:53:23< MoniqueMitchell> Me too. :)
20:53:52< pierregiraud> I’d be happy to learn the methods you’ll be using for the design process
20:54:06< pierregiraud> will you use balsamiq for the wireframes as well?
20:54:38< MoniqueMitchell> We had some issues with the share functionality for balsamiq.
20:55:21< MoniqueMitchell> As such , the final site is set up like a normal website (I'm not a webdesigner but I am assuming html/C++)
20:55:21< pierregiraud> are you also doing some usability tests?
20:55:33< dodobas> there is mozilla Pencil project... but that's desktop
20:55:38< MoniqueMitchell> sorry *CSS
20:55:49< dodobas>
20:55:56< pierregiraud> dodobas: this works well too
20:56:10< pierregiraud> both pencil and balsamiq have interesting features
20:56:23-!- harry-wood [~Adium@] has joined #hot
20:57:12< dodobas> does anyone have anything else to report... other then harry-wood, of course, ... :)
20:57:58< dodobas> AndrewBuck: do you want to share anything?
20:58:03< MoniqueMitchell> Unfortunately, the mock-up is fairly static. The clickable features are to get an idea of what we're suggesting. The "website" is primarily a way to visualize our ideas. The project was designed to take the first part of a design process in mind: feedback from the user->design ideas
20:58:10-!- sev_hotosm2 [] has joined #hot
20:58:59< pierregiraud> MoniqueMitchell: sorry I couldn’t attend the meeting last thursday
20:59:05< AndrewBuck> Well, I am looking for subject matter experts to help me with my ontology work, also I would like to speak with all of the stylesheet and routing engine designers about making use of the ontology for their work.
20:59:15< MoniqueMitchell> It will stop here for most of the teammembers due to the nature of the course, but it would be ready for the next cycle at this point (highfidelity wireframes->user testing)
20:59:24< AndrewBuck> But aside from that I can't think of anything right offhand.
20:59:43< AndrewBuck> ptressel: are you still aournd?
20:59:53< ptressel> Yep
* MoniqueMitchell sits down 21:00
21:00:18< pierregiraud> MoniqueMitchell: it’s nice to hear that students are learning how to design applications
21:00:40< pierregiraud> we (open source developers) are usually bad on this topic
21:00:45< AndrewBuck> ptressel: did you want to discuss the todo plugin thing now, or after the meeting?
21:00:52< pierregiraud> thanks MoniqueMitchell
21:01:10< MoniqueMitchell> my pleasure!
21:01:32< AndrewBuck> yeah, sorry everyone, I didn't mean to interrupt, I saw dodobas's message and thought we were moving on.
21:01:34< pierregiraud> dodobas: I think we’re done with the tasking mangaer
21:01:36-!- mkl [~mkl@] has joined #hot
21:01:39< ptressel> AndrewBuck, Just didn't want to interrupt.
21:01:58< pierregiraud> that’s OK guys we were actually finished
21:02:04< AndrewBuck> ok, cool
21:02:11< ptressel> ;)
21:02:34< AndrewBuck> so ptressel was thinking of adding a way in josm to more systematically scan over an are like a tm task tile.
21:02:48-!- cmpera [~cmpera@] has joined #hot
21:02:49< AndrewBuck> I was suggesting that this be added to the todo plugin in josm.
21:02:52< ptressel> Don't hit me but...I may have something a bit different in mind from the todo plugin.
21:03:05< pierregiraud> what does « scan over » means?
21:03:06< AndrewBuck> yes, go ahead.
21:03:08< ptressel> Can I describe the need?
21:03:42< AndrewBuck> pierregiraud: to look over an area, for xample to see if there are missing buildings, etc. and make sure you covered the whole area at least once.
21:03:50< ptressel> I often find, when working on a HOT task defined by a large rectangle that I loose place.
21:04:22< ptressel> I am trying to cover the entire region, so try carefully to step up and down or side to side in stripes.
21:04:38< ptressel> But when I'm chasing features, I get off line.
21:04:59-!- harry-wood [~Adium@] has quit [Quit: Leaving.]
21:05:02< ptressel> And it's not necessarily easy to tell what I've covered by added objects.
21:05:23< AndrewBuck> ptressel: yeah, that is exactly what I was thinking of.
21:05:24< ptressel> If I commit often, I no longer have uncommitted objects to tell what I did.
21:05:32< mkl> you could use filters to show only newly added objects
21:05:39< ptressel> ^^^
21:05:46< ptressel> I've been burnt by that.
21:05:53< ptressel> I now commit **often**.
21:05:59< ptressel> Do not like doing merges.
21:06:16< ptressel> I might be working on one task for hours.
21:06:21-!- harry-wood [~Adium@] has joined #hot
21:06:36< ptressel> And if it's an activation, others will be working on adjacent rectangles.
21:06:44-!- robertsoden [] has quit [Quit: robertsoden]
21:07:00< AndrewBuck> my suggestion for the todo list was to add another kind of list that covers an 'area' rather than a list of node/ways/relations. To add an area you would zoom out so your josm map view covers the area you want to map and then click a button in the task manager...
21:07:01< mkl> or everything created after the start of your session
21:07:06< ptressel> In any case, I don't want to have to keep track of what regions I've covered by hand.
21:07:08-!- jgc [~jgc@] has joined #hot
21:07:25< ptressel> AndrewBuck, That's what I meant by "different". ;-)
21:07:54< ptressel> No adding areas -- when one activates the plugin, it splits the current task into a grid.
21:07:56< AndrewBuck> then you zoom in to the level where you can see the features you want to work with and press a second button. Then the task manager fills the area with "screens" which are the zoom level you asked for and cover the wwhole region.
21:08:20< ptressel> The grid size is based on the viewport size and zoom level.
21:08:33< AndrewBuck> yeah, we are describing the same thing.
21:08:43< ptressel> One uses (e.g.) arrow keys to go to the "next" unworked grid.
21:09:23< ptressel> A small window w/ the grid shows which have been done.
21:10:05< AndrewBuck> that would be cool.
21:10:34< ptressel> The main difference may be that there is not set of elements added to a todo list.
21:10:53< AndrewBuck> well, the elements I was talking about were the little square you are talking about.3
21:11:19< AndrewBuck> so you mark each "square" complete and then eventually you have the whole list of them done and the whole area has been covered.
21:11:22< ptressel> Ok, but the user doesn't have to add them -- that's automatic. And it's not a list, but a grid.
21:11:40< ptressel> Not a list...array.
21:11:52< AndrewBuck> Well, to add them you would just zoom in once and press the add button and then it populates the whole area with them.
21:12:13< AndrewBuck> we are describing the same thing.
21:12:16< ptressel> The other thing is... When I move to a new section, I want it to overlap with the previous one.
21:12:31< ptressel> So the grid elements are smaller than the viewport.
21:12:46< AndrewBuck> yeah, Iw as thinking there would be something like a slider bar for what percent of the view to overlap.
21:12:58< ptressel> Stepping to the next moves over by some fraction of the grid size.
21:13:07< AndrewBuck> so you set that at the beginning when the list of "squares" is created.
21:13:26< ptressel> But defaulted to something reasonable.
21:13:31< AndrewBuck> of course.
21:14:29< AndrewBuck> yeah, so then yuo press a button to mark a square complete and then use an arrow key to move to a neighboring one.
21:14:34< ptressel> Marking as complete would (for preference) also be a shortcut.
21:14:36< ptressel> Yes
21:14:45-!- harry-wood [~Adium@] has quit [Quit: Leaving.]
21:15:14< ptressel> Actually, I was thinking of having "mark as complete" just move on to the next.
21:15:31< pierregiraud> will the users still remember to go back to the tasking manager when the TM task is complete?
21:15:32-!- MoniqueMitchell [] has quit [Quit: CGI:IRC]
21:15:54< AndrewBuck> Yeah, like I said, my suggestion would be to add that to the todo plugin. I had planned to add a 'tab' menu across the top with the first tab being the one that is there now, and then there would be one for what you are describing, and also tabs for showing the completed list, etc.
21:15:54< ptressel> Not sure of the spelling but...boustrophedon ?
21:16:07< ptressel> "like an ox plowing a field"
21:16:47< AndrewBuck> pierregiraud: yeah, I think so, this is just a tool to help you keep track of your progress in larger tm tiles.
21:16:57< ptressel> There were a bunch of other possible features, like time reminders or # addition reminders for committing.
21:17:05< AndrewBuck> ptressel: yeah, you work your way across and then reverse direction for the next row
21:17:10< ptressel> Right.
21:17:27< pierregiraud> what if the tasking manager tiles were smaller?
21:17:39-!- clara-afk [] has quit [Ping timeout: 480 seconds]
21:17:40< ptressel> Even little ones are big. ;-)
21:17:41< pierregiraud> stupid question perharps
* pierregiraud doesn’t do mapping much 21:18
21:18:03< AndrewBuck> pierregiraud: for things like roads you want big tiles but then it gets a bit tricky to track exactly where in the tile you have done.
21:18:03< ptressel> The grid elements are < size of the viewport.
21:18:26< ptressel> If I'm drawing buildings, I'm usually zoomed in fairly closely.
21:18:44< pierregiraud> I see
21:18:49< AndrewBuck> it is also really helpful even on smaller tm tiles just to scan through and look for things like missed buildings.
21:19:32< AndrewBuck> so yeah, sounds like we are in agreement then on the basic design of the thing.
21:19:37< pierregiraud> validators will also like this feature
21:19:39< ptressel> When working on Carles peninsula, the TM squares had about 20-30 steps across, IIRC.
21:19:51< ptressel> :D
21:20:14< AndrewBuck> yeah, which means 20 across and 20 down = 400 smaller "screens" to look at for a single tm job.
21:20:36< ptressel> So, AndrewBuck, the *real* question you want other hands on this?  ;-)
21:20:48< PovAddictW> AndrewBuck: do you think kuona will be needed again in the future?
21:21:08< AndrewBuck> ptressel: I could definitely use the help. My coding skill are very limited.
21:21:31< AndrewBuck> PovAddictW: There was some talk recently of making something like kuona again.
21:21:40< AndrewBuck> this is a bit different though.
21:22:00< ptressel> I used to work on Java desktop apps, but haven't recently except for Android, which does not count. ;-)
21:22:27< AndrewBuck> yeah, if you are interested in coding this we should talk on skype or mumble.
21:22:45< ptressel> Ok. I just installed mumble last night.
21:22:49< AndrewBuck> do you think it makes sense to add it to the todo plugin on a tab like I suggested?
21:22:56< PovAddictW> if you're good with Java desktop code, you can help with JOSM plugins
21:23:11< AndrewBuck> I figured it is basically the same thing just a different kind of a list.
21:24:09< ptressel> I was trying to think of a way that one would not need to "materialize" the grid, as it's quadratic size in the number of steps across.
21:24:51< ptressel> But it's easier if there are todo items for each grid element -- that lets them be marked complete more simply.
21:25:00< AndrewBuck> yeah, I like your idea of showing it graphically, my idea was to use an extra layer like the validator does. I wanted to add a layer like that to the todo plugin anyway but my coding skill are just not up to it.
21:25:52< AndrewBuck> We can discuss details like this on voice though, anything else for the workgroup meeting?
21:26:00< ptressel> I was thinking a layer would be nice in addition. Toggle on and off a transparent layer showing what is done.
21:26:07< AndrewBuck> exactly
21:26:08-!- sev_hotosm2 [] has quit [Ping timeout: 480 seconds]
21:26:41< ptressel> I'm glad you're using git for the todo plugin.
21:27:03< AndrewBuck> yeah, much more amenable to having others work on the code.
21:27:08< ptressel> Collaboration with svn is ... is ... having trouble coming up with a polite word.
21:27:15< ptressel> A bear.
21:28:03< AndrewBuck> Well, I think we are in agreement then about the design.
21:28:11< ptressel> ;-)
21:28:46< ptressel> I was mainly worried that the grid elements would get added to a textual list and the user would have to select them.
21:28:47< AndrewBuck> Does anyone have anything else for the meeting?
21:29:09< dodobas> if not... thank you all for attending
21:29:22< pierregiraud> thanks
21:29:31< ptressel> o/
21:29:45< dodobas> I'll upload the log and create TLDR for this meeting...
21:29:59< dodobas> next one should be in 4 week
21:30:06< AndrewBuck> thanks dodobas, good meeting.
21:30:21< dodobas> June 2nd
21:30:36< dodobas> I'll announce it on time...