Humanitarian OSM Team/Working groups/Technical/meeting 2013-10-21

From OpenStreetMap Wiki
Jump to: navigation, search

Meeting to discuss Tech topics on Monday 21st October 2013

IRC log:

18:56:17< dodobas> everyone gettign ready ? :)
18:57:39< ybon> o/
19:01:24< dodobas> hello everyone... we are about to start Tech WG meeting
19:02:19< dodobas> does anyone want to start... alexbarth, f00bar, Haaninjom, harry-wood1, pierzen, pnorman, ybon ?
19:03:04< ybon> well, honnestly, on my side, nothing very relevant, I'm doing scuba diving in Java one more week :/
19:03:31< dodobas> and that's a bad thing ? :)
19:03:35< pierzen> my first chimney fire today.
19:04:27< harry-wood1> is that a technical term for a type of bug?
19:04:36< dodobas> pierzen: oh... is everything fine ?
19:04:41< pierzen> buggy-wood
19:04:49< pierzen> dodobas: yes
19:05:11< dodobas> good
19:05:38< harry-wood1> also is "scuba diving in java" a technical term for a type of experimentation with the language?
19:05:48< ybon> dodobas: the Java I'm talking about is not an island, actually :p
* chippy lurking nothing to report though 19:06
19:06:24< dodobas> harry-wood1: probably ybon wants to rewrite Linus Torvalds scuba app that's in C on github :)
19:06:44< chippy> oh, I've one thing to report
19:06:54< dodobas> chippy: please do
19:07:43< harry-wood1> I was playing with vagrant at the weekend. Not much of an expert at such things, but I did come across geobox: and also vagrant-webmaps:
19:08:03< harry-wood1> nifty. Makes me think that this would be the way to go for preparing HOT USB key
19:08:10< chippy> the SDS Separate Data Store test server was on my free Amazon ec2 instance - it's run out, and they are starting to charge me - so it could be a good candidate to move elsewhere. It's not much per month so no urgency
19:08:37< chippy> I think folks were using it for training / familiarity purposes
19:08:50< dodobas> chippy: ok... ill move it to amelia...
19:09:43< chippy> ok will contact anon, dodobas :)
19:09:56< dodobas> chippy: anon ? :)
19:10:00< chippy> another time
19:10:06< dodobas> ahh, k
19:10:46< harry-wood1> chippy I was looking at this list: we have separate wiki pages on a some tools, but no wiki pages about 'Separate Data Store' or 'Visual Tag Chooser'
19:10:53< dodobas> harry-wood1: well, could be.. we need to minimize the network bandwidth requirement for filed depolyments
19:11:01< ybon> ( dodobas: héhé, I'm actually doing autocompletion + fuzzy search on OSM data with ElasticSearch ;) )
19:11:52< chippy> harry-wood1, good idea - will add in some pages
19:12:37< harry-wood1> dodobas: yeah that's true. Also vagrant requires install of virtualbox obviously, so it's lightweight in some ways but not in others. I guess just the general principle of scripting the install of stuff… is the way to go for building a HOT USB key
19:13:17< harry-wood1> geobox doesn't work because some of the URLs no longer exist… so there's the disadvantage right there :-)
19:13:26< dodobas> hehe
19:14:06< dodobas> if no one wants to report anything... i can share what i've done
19:15:42< dodobas> basically... i've manged to deploy tasking-manager-V2 that pierregiraud develops
19:15:52< dodobas> give me a few seconds... ill start it up
19:16:27< dodobas> there are some issues with the deployment process... so i need to create some pull requests
19:16:43< dodobas> and iron a few glitches....
19:18:02< dodobas> there was some disscussion about organizing a coding marathon for tasksV2...
19:18:52< dodobas> also... frederik added nepal and mongolia to the export server... and provided documentation on how to actually do it
19:19:06< dodobas> (I still need to check to docs)
19:20:05< dodobas> we also got a domain hosted on the amelia... which makes it easy to add a subdomain for any purpose
19:20:32< dodobas> like which is a new Drupal7 site
19:20:33< skorasaurus> now here although late
19:20:43< dodobas> skorasaurus: ok... you can go after me...
19:20:58< dodobas> I've also moved tasksV1 to the ameila...
19:21:11< dodobas>]
19:22:00< dodobas> the idea is to move it to the ameila... so if you can test it out, we would be grateful
19:22:35< dodobas> but... there is a itsybitsy problem with the current depolyment...
19:23:06< skorasaurus> sorry for jumping in late, and maybe you already explained it.
19:23:11< dodobas> tasksV1 use sqlite database for production...
19:23:13< skorasaurus> but why move tasksv1 to amelia ?
19:23:46< skorasaurus> was it using too many resources on other servers before  ?
19:23:48< dodobas> skorasaurus: reduce costs... amazon is expensive... you can discuss
19:23:57< dodobas> more with wonderchook...
19:24:05< skorasaurus> ah, gotcha.
19:24:25< skorasaurus> didn't know it was on there. thought it was on another osm server.
19:24:35< skorasaurus> no problem then :)
19:24:44< dodobas> so... sqlite DB means that you can use only a single process to access the DB...
19:25:42< dodobas> i would like to move it to a proper DB (postgresql) but... ill probably need to manually create/rewrite the dump... as SQLIte is not compatible with pgsqlseconds]
19:27:17< dodobas> so.. that's about it... skorasaurus you can continue... while i prepare tasksv2 instance
19:29:11< skorasaurus> not much on the front w/ me.
19:29:27< skorasaurus> HDM-Cartocss is continuing on, has been deployed worldwide.
19:30:01< skorasaurus> ybon fixing a small rendering bugs within past couple weeks.
19:30:41< skorasaurus> I presented on the HDM/CartoCSS rendering at NACIS, a relatively well-known cartography conference 2 weeks ago.
19:30:51< skorasaurus> the map was well received.
19:30:56< skorasaurus> and received many compliments.
19:31:31< ybon> yay :)
19:31:37< skorasaurus> although I learned that it can be difficult to see roads if you are projecting it on a screen in a bright room.
19:31:45< skorasaurus> although that wasn't the intention of our map ;)
19:32:05< ybon> hehe
19:32:19< skorasaurus> yep :)
19:33:00< skorasaurus> one thing to note - this wasn't a HOT thing necessarily, but with OSM at large, whether to map domestic violence shelters.
19:33:16< skorasaurus> there's a very very small number (no more than 30 or so) in the planet.
19:33:58< skorasaurus> (there's 2 tags or tag combinations that I Know of for them.
19:34:10< pnorman> I'm pretty sure there's more than that. I'd expect 20+ in Canada alone
19:34:58< pierzen> I am not sure that domestic violence shelters wants to be mapped very precisely.
19:35:06< skorasaurus> amenity=social_facility + social_facility:for=abused
19:35:14< skorasaurus> pierzen agreed.
19:35:43< pierzen> I friend of mine did everything for people to not know where they are.
19:35:57< skorasaurus> and another tag that kate found (this topic was discussed on twitter last week)
19:36:38< skorasaurus> pierzen, yes, I have learned that in most cases, those types of shelters do not wish to have their geographic locations available to the public.
19:37:10< skorasaurus> specifically shelters that treat victims of sexual and domestic abuse.
19:38:04< skorasaurus> there's also - another tag that kate knew of.
19:38:31< pnorman> looking for a specific question that relates, sec
19:39:25< skorasaurus> pnorman: there's many many more facilities of that alone in canada, and all of the world, i agree, although there's very few in the OSM db at the moment. and we should highlight and think of approaches so that they are not included in OSM data.
19:39:33< dodobas> ( tasksv2 @ )
19:39:59< pnorman> is probably the most relevant. anyways, this isn't really a HOT question, but a wider question of mapping facilities that do not wish to be mappedquit [Remote host closed the connection]
19:41:12< pierzen> dodobas: got it
19:41:39< skorasaurus> pnorman, agreed.
19:41:45< skorasaurus> it's a question at large.
* pnorman has reverted facilities removing data about themselves from the map locally; unrelated to shelters of this type 19:42
19:46:10< dodobas> does anyone wnat to add something
* pnorman only caught the end 19:46
* ybon ok 19:46
19:47:16< dodobas> if not, we are going to end this meeting and reschedule for 4.11
19:47:54< dodobas> i'll add the log to the wiki, and create a TLDR summary like the laste time
19:47:56< dodobas> *last
19:48:19< dodobas> if you can test tasksv1 and taskv2...
19:48:50< dodobas> thank you for comments and reports
19:49:10< skorasaurus> dodobas: for tasksv2, should we create new tasks ?
19:49:19< skorasaurus> or just map the existing ones.
19:50:24< skorasaurus> and bug reports under the github repo ?
19:50:33< dodobas> skorasaurus: well, pierregiraud will have more info
19:50:58< dodobas> however now would be the perfect time to break it :)
19:51:26< dodobas> skorasaurus: the issues can be found at..
19:53:27< pnorman> fwiw, googling for "domestic violence shelter vancouver" pops up 3 pins on a google map immediately
19:54:07< skorasaurus> dodobas: thanks :)