Humanitarian OSM Team/Working groups/Technical/meeting 2013-12-09

From OpenStreetMap Wiki
Jump to navigation Jump to search

Meeting to discuss Tech topics on Monday 9th December 2013

IRC log:

18:01:30< dodobas> we are about to start HOT Tech WG meeting ... if anyone is around :/
18:01:33-!- FredB [~makina@mir31-1-82-224-13-158.fbx.proxad.net] has quit [Quit: FredB]
18:01:41< russdeffner> I am here
18:01:57< dan4dm_> hi
18:03:36< pgiraud> I would be happy to attend
18:03:39-!- Irssi: #hot: Total of 44 nicks [0 ops, 0 halfops, 0 voices, 44 normal]
18:03:47< pgiraud> but I have to ride my bike home now
18:04:10< dodobas> pgiraud: so you can start if you want :)
18:04:28< pgiraud> ok I have only 5 minutes
18:04:35< dodobas> type fast :)
18:04:54< pgiraud> I just wanted to say that I really want to merge the workflow branch into master
18:05:05< pgiraud> I got support from Daniele
18:05:19< pgiraud> and I got feedback from several users
18:05:26< dodobas> Daniele ?
18:05:49< pgiraud> Daniele Venzano
18:05:57< pgiraud> he helped fixing some issues
18:05:58< dan4dm_> pgiraud: is the "workflow" branch the one on tasks2.hotosm.org?
18:06:02< dodobas> you are taking about tasking manager v1 ? or v2 ?
18:06:03< pgiraud> exactly
18:06:13< pgiraud> tasking manager v1 branch v2
18:06:17< pgiraud> oops
18:06:18-!- gregory [~gregory@host86-148-33-3.range86-148.btcentralplus.com] has quit [Ping timeout: 480 seconds]
18:06:22< pgiraud> v1 branch workflwo
18:06:35< pgiraud> this is the one installed on tasks2.hotosm.org
18:06:59< dodobas> ok
18:07:02< pgiraud> as well as on http://tasks-dev.hotosm.org/
18:07:20< pgiraud> people got confused on the several we have now
18:07:47< pgiraud> we really need to clear things up
18:07:56< dodobas> we will need to conslidate... but... we need to fix deployment
18:08:19< pgiraud> I agree
18:08:48< pgiraud> I also had question from a guy who wanted to create a puppet module for the tasking manager
18:09:33< dodobas> he was on IRC, right? i seem to remeber that
18:09:49< pgiraud> yep, he asked here on IRC
18:10:23< pgiraud> I should be able to find his name since he open an other issue on github
18:10:45< dodobas> pgiraud: what's stopping you to do the merge ?
18:10:46< pgiraud> he also asked if it would be possible to have the tasking manager available as a package on pypi
18:11:09< pgiraud> well some feedbacks make me think that some users won't be completely happy
18:11:38< pgiraud> I still have to explain why users now need to click several times to lock a tile
18:12:06< pgiraud> but, well, I'm happy with this new branch and I should definitely merge
18:12:12< pgiraud> it has a lot of improvements
18:12:24< pgiraud> the remaining issues can be fixed later
18:12:26< dan4dm_> pgiraud: I am the person who put a github issue up with feedback on that. I suspect that the multi-click is related to the direct-linking-to-tasks?
18:12:40< pgiraud> exact
18:12:48< pgiraud> I call it read-only link
* xamanu is late but present now 18:12
* xamanu has read the log of the meeting till now 18:13
18:13:39< xamanu> Is the meeting still going on?
18:13:40< dodobas> pgiraud: well, you can ask on the mailing list, or someone from HOT
18:13:49< dan4dm_> I'd suggest that a GUI click-on-a-square could still lock, even if the URL direct link does not. But I won't go on and on, since I already said it on github!
18:13:57< dan4dm_> xamanu: yes
18:14:02< xamanu> nice :)
18:15:00< dodobas> xamanu: pgiraud has the microphone at the moment, when he finishes... you can continue :)
18:15:00< pgiraud> I still think that people may want to see the history for a tile by clicking on it
18:15:24< pgiraud> even if they don't want to lock it
18:15:48< pgiraud> I have to leave now, I'm done, next
18:15:59< dodobas> pgiraud: the load on the servers is lower at the moment ?
18:16:02< dan4dm_> I guess it comes down to: do we (you) expect to have more of a problem with people working on tasks without locking them, or with people locking tasks without working on them!
18:16:46< dan4dm_> ah, sorry for taking from pgiraud's limited time
18:16:57< pgiraud> I'm really open to suggestion: I'll send an email to discuss this on the mailing list (with screenshots)
18:17:35< dodobas> AndrewBuck, harry-wood, mkl, ybon, aynone around ?
18:17:42< AndrewBuck> hey
18:17:52< dodobas> xamanu: do you have anything to report ?
18:18:13< xamanu> yes, I would like to speak about the hot website
18:18:27< harry-wood> some activity on the hot website lately I see :-)
18:18:43< harry-wood> (github drupal 7 repo)
18:18:45< dodobas> ok... so the order is xamanu, AndrewBuck, harry-wood ... :)
18:18:59< xamanu> I may introduce myself, as I haven't attended any HOT tech meeting yet:
18:19:30< xamanu> I'm Felix from Germany living in Nicaragua for some years now. Check http://felix.delattre.de for my website.
18:19:52< xamanu> icaragua for some years now. Check http://felix.delattre.de for my website. I came to HOT as I was invited to speak at SOTM-US in San Francisco about out Managua (Nicaragua's capital) mapping effort on transit mapping.
18:20:11< xamanu> sorry for the bad copy and paste :)
* dodobas waves 18:20
18:20:22< dan4dm_> hi felix
18:20:28< xamanu> Hi all :)
18:20:39< xamanu> Anyway, in SF I met Mikel who asks me to help with the Drupal port, as I'm a Drupal dev mainly
18:20:52< xamanu> https://drupal.org/user/359937
18:21:13< xamanu> So I worked together with clara, who as I think has attended some meetings before, on the HOT website
18:21:56< xamanu> And, we are somehow at the point as we would like to have you checking the new version of the website: http://www.dev.hot.org
18:22:39< xamanu> We propose to get this live pretty soon
18:22:51< dodobas> http://www.dev.hotosm.org
18:23:01< dodobas> is the correct link
18:23:11-!- natesmith [~natesmith@50-198-132-58-static.hfc.comcastbusiness.net] has joined #hot
18:23:15< xamanu> thanks dodobas :)
18:23:43< xamanu> I think the site is not in a perfect state, but it is as good, or slightly better than the existing site
18:24:02< xamanu> and getting more content added to the old site is anoying as this has to be migrated by hand to the new one
18:24:31< xamanu> so, I would like to invite you to test, so we can move forward and then keep on fixing and improving the site running drupal version 7
18:25:53< dan4dm_> xamanu: one small question. The "skip to main content" link (visible when css is off) - it jumps to the newsfeed, not to the featured project - that may be unintended (the "featured project" seems like main content to me!)
18:26:33-!- pgiraud [~pgiraud@115.19.0.109.rev.sfr.net] has quit [Read error: Operation timed out]
18:26:33< dodobas> xamanu: ok, kuul, do you have anything to add, if not, i'll just mention some of the technical issues we should solve before going to production
18:26:38< xamanu> dan4dm_ good point. I can open an issue for that and fix it. btw, we use the issue tracker at: https://github.com/hotosm/hotosm
18:27:21< xamanu> dodobas, fine for me
18:27:39< dodobas> ok...
18:28:35< dodobas> AFAIK, HOT still does not have anykind of SSL certificate, at least an official one
18:29:51< ybon> ( o/ hi, sorry, I'm in Ndjamena with a poooooor Internet and basically nothing special in my done list :( )
18:29:59< dodobas> I did have a brief discussion withour executive director, but we did not purchase one, at least not yet
18:31:00< xamanu> ok. do we need tha SSL certificate to use the server and the d7 hot website?
18:31:24< dodobas> no... but all of the passords are going to be snt over the wire unencrypted
18:31:30< xamanu> I agree that this would be nice to have! But should not be a show stopper
18:31:34< dodobas> +passords
18:31:42< dodobas> damn.. :)
18:31:43< xamanu> dodobas: at it is right now...
18:31:50< xamanu> as it is
18:31:54-!- FredB [~makina@2a01:e35:8b19:f610:a8e5:8529:2919:c27d] has joined #hot
18:32:06< dodobas> there are no show stoppers
18:32:39< xamanu> dodobas: we could use a https version when logging in (admins only) and use a CAcert signed certificate for now...
18:32:53< xamanu> but usual user would get served a non-ssl version of the website
18:33:11< dodobas> however, we should discuss what are we going to do about the dev instance
18:33:54< dodobas> do you want me to prepare a clean production instance, or move dev -> production and prepare a clean dev instance ?
18:34:26< xamanu> dodobas: either one would be fine for us, i guess
18:34:36< dodobas> thats fine :)
18:34:51< xamanu> I prefer using our infrastructure, than using the Pantheon stuff we are using right now
18:35:08< dodobas> i just wanted to raise my concerns
18:35:27< xamanu> right now for our workflow, pantheon has been our dev and the server's dev our drupal 7 live instance.
18:35:27< dodobas> yes it will be hosted on the new server...
18:35:59< xamanu> \o/
18:36:25< dodobas> xamanu: we can discuss abot the details when we get the green light to switch
18:36:37< xamanu> dodobas: of course
18:36:39< dodobas> AndrewBuck: grab the mic...
18:36:45< xamanu> wait pls
18:36:53< AndrewBuck> I don't really have anything to add anyway
18:37:08< xamanu> what is the next step?? I'm not familiar with our communication workflow
18:37:27< xamanu> But I want to get this live as soon as possible
18:37:46< xamanu> Can we get to an agreement?
18:37:49-!- mkl [~mikel@c-68-50-131-97.hsd1.dc.comcast.net] has quit [Ping timeout: 480 seconds]
18:38:01< dodobas> xamanu: well, wait for comments, report that no one had any mayor complaints
18:38:02< xamanu> What do we need to get this live?
18:38:50< xamanu> dodobas: what would be a good time frame for that? One week?
18:38:55< dodobas> and then we will start discussing when to switch over
18:39:02< xamanu> Should I send an email to the mailing list, i guess?
18:39:58< dan4dm_> hasn't xamanu already asked the mailing list for feedback?
18:40:01< xamanu> dodobas: is one week ok? after sending it to the list. can we discuss this personally between us two, or do we need another HOT tech meeting?
18:40:07< dodobas> xamanu: at the moment, i would say... either this week, or sometime in early january
18:40:37< dodobas> xamanu: wonderchook, our executive director,, should make the final decision
18:40:51< xamanu> dodobas: ok. Please let us do this anytime soon (if there are no mayor complaints). As I don't like migrating content and comments by hand
18:41:48< dodobas> xamanu: it's not my decision :)
18:42:00< dodobas> im just a server admin
18:42:47< xamanu> ok. i'll write an email to the list. as I haven't before, as dan4dm_ said. Maybe clara did!?
18:43:00< dan4dm_> ah yes it was clara
18:43:43-!- mkl [~mikel@c-68-50-131-97.hsd1.dc.comcast.net] has joined #hot
18:43:46< xamanu> dan4dm_ I saw it. thanks. I will write in response to her. Getting things pushed forward
18:44:01< xamanu> thanks, dodobas and dan4dm_ for answering my questions
18:44:02< dodobas> ok, so, report any comments you have, explain the situation, and ask for more testing
18:44:05< xamanu> I'm done... :)
18:44:21< dodobas> im gong to discuss with wonderchook in the morning...
18:44:38< dodobas> as it's 0:43 in Jakarta ... :)
18:44:59< xamanu> ok, thanks dodobas!
18:45:19< dodobas> AndrewBuck: nothing like.. mumble server, tasks2, any issues with the hot infrastructure ?
18:45:51< AndrewBuck> The mumble server is working fine for me although the folks in dakar were having trouble connecting.
18:45:58< AndrewBuck> That could be on their end though.
18:46:38< AndrewBuck> As for tasks2 it worked well, i haven't used it much myself though so nothing to add other than my previous discussions with you and Pierre.
18:47:00< dodobas> AndrewBuck: can they connect to som other mumble server ?
18:47:05< dodobas> *some
18:47:19< AndrewBuck> Not sure, I just got an email from Pierre Beland about it
18:47:36< AndrewBuck> Their internet is pretty shaky though so it is not surprising they would have issues.
18:48:00< AndrewBuck> I expect they will try again, I also might have them try tcp mode instead of udp if the problem persists
18:48:24< AndrewBuck> I am connected to the server now though and have been for days with no disconnects so it doesn't seem to be a problem there.
18:48:52< dodobas> yeah...
18:49:31< dodobas> AndrewBuck: if that is all, pass the totem to harry-wood
18:49:50< AndrewBuck> Ahh yes, sorry
18:49:54< AndrewBuck> Go ahead harry-wood
18:49:57< harry-wood> I am mega busy at work today
18:50:31< harry-wood> but… I get an email from Andrew Braye of the British Red Cross asking about things for developers to work on
18:50:43< harry-wood> he had some volunteers with PostGIS skills
18:50:56< harry-wood> so I sent quite a long reply giving some ideas
18:51:06< harry-wood> I should share that reply somewhere really
18:51:30< dodobas> like 'how to get involved page' on the website ? :)
18:51:57< harry-wood> yeah so he'd found that page already and was asking if that was the best thing to point people at
18:52:19< harry-wood> to which I said "yes! ….although here's some other ideas"
18:52:39< harry-wood> I pointed him at our HOT technical working group page on the wiki
18:52:48-!- alexbarth [~alexbarth@50-198-132-58-static.hfc.comcastbusiness.net] has quit [Remote host closed the connection]
18:52:50< harry-wood> which has some of the same things listed… plus a few others
18:53:01< dodobas> ok kuul
18:53:15< harry-wood> good thing if we can keep that page up to date and tidy
18:53:31< harry-wood> (needs a bit of work at the moment)
18:53:58< AndrewBuck> one thing they could work on is helping orgs like themselves and others get their own local databases set up with the planet file as well as overpass and such.
18:54:29< AndrewBuck> Not development work per-se but if they know how all that works they could really help out a lot of people.
18:54:59< harry-wood> yes. I suggested trying out installing tilemill and getting the humanitarian style working
18:55:19-!- tmcw [~tmcw@50-198-132-58-static.hfc.comcastbusiness.net] has quit [Remote host closed the connection]
18:55:20< harry-wood> involves a certain amount of postGISing I suppose
18:56:25< harry-wood> install overpass API locally, I hadn't particular thought about, but yes I guess that could be an interesting challenge :-)
18:57:21< dodobas> harry-wood: if you have finished, anyone else wants to report something, dan4dm_ ?
18:57:48< dan4dm_> ok just briefly
18:58:13< dan4dm_> I was at the London hack weekend recently - me and one other person had a go at installing Task Manager and doing a bit of hacking on it
18:58:44< dan4dm_> we discussed whether it should be on postgres rather than sqlite (for concurrency) but we didn't get near anything as tricksy as that in the end
18:59:08< dodobas> (it shuld be, but it's not :/)
18:59:25< dan4dm_> I pushed a couple of tweaks as feature requests... but it turned out that they were irrelevant because pgiraud's "workflow" branch already had them in!
18:59:53< dan4dm_> so that's just unfortunate timing on our part - if we'd known, we would have started with that branch. but at least we know we can hack on it
19:00:25< dan4dm_> so, my feedback on tasks2 etc is already out in the world. once the branch is merged I will probably make some suggestions in the form of pull requests.
19:00:28< dan4dm_> but that's all from me.
19:00:59< dodobas> ok, tnx, it's good to see more people interested in tasking manager development
19:01:17< dan4dm_> oh though I may as well ask: is postgres port an explicit aim, plans in the works? or just an idea?
19:01:38< dan4dm_> rob mentioned specific postgres features that would be good such as row-level locking
19:02:30< dodobas> dan4dm_: well, for tasking v1... it might not be more then an idea...
19:02:45< dodobas> as tasking v2 already supports postgres
19:03:23< dan4dm_> brill
19:03:44< dodobas> and even with postgres, you are still going to have 'problems' when creating new jobs
19:04:17< dan4dm_> and there is currently active development on v1 workflow branch - how will that relate to v2?
19:04:48< dodobas> i was making suggestions on the line of - improving job creation code
19:05:31< dan4dm_> what I mean is: will the current work on v1 be "lost" if/when v2 comes?
19:05:41< dodobas> dan4dm_: its hard to tell, v2 has diverged from v1, i think it's almost a complete rewrite
19:06:13< dodobas> and there were some disscusion on getting funding for v2 development
19:06:27< dodobas> however, the pace is slow...
19:06:38< dan4dm_> ok thanks
19:06:53< dodobas> pgiraud willknow more...
19:07:39< dodobas> so... i just wanted to say that we have a couple of new services
19:08:12< dodobas> a Mumble/Murmur server -> talk.hotosm.org
19:09:09< dodobas> and a maposmatic service for philippines with HDM - hot style
19:09:23< dodobas> http://maposmatic.dev.hotosm.org/
19:10:11< dodobas> oh yeah... and http://tasks2.hotosm.org
19:11:56-!- jaakkoh [~jaakkoh@186.77.205.121] has quit [Ping timeout: 480 seconds]
19:12:09< dodobas> to be honest, i did not check the content of tech WG wiki page, other then updating meeting logs
19:12:14< dan4dm_> \o/
19:12:46< dan4dm_> tho maposmatic server says its db+daemon aren't running
19:13:44< dodobas> dan4dm_: yeah... not sure why... but it works for the philippines area (there is only data for that area)
19:14:41< dodobas> if no one else has anything to add... we are going to end this meeting
19:15:14< dodobas> ill post the logs on the wiki, and create a short summary
19:15:29< dodobas> im off to bed :)