14:30:51 #startmeeting metrics team 14:30:51 Meeting started Thu Jul 19 14:30:51 2018 UTC. The chair is karsten. Information about MeetBot at http://wiki.debian.org/MeetBot. 14:30:51 Useful Commands: #action #agreed #help #info #idea #link #topic. 14:31:06 https://storm.torproject.org/shared/5h1Goax5eNusxjXJ_Ty5Wl7hFR1uqCReUiN8xdlBG8T <- agenda pad 14:31:30 i just added the one topic 14:32:54 I added a few more. 14:32:59 * Priority for #25625? (irl) 14:33:06 i've been working this week on the ooni vanilla tor inclusion on tor metrics, i saw your comment on #25625 and wondered if perhaps this was sponsor 13 related and should be prioritised over the vanilla tor task? 14:33:24 ah, no, that is not the case. 14:33:38 no need to interrupt that other task. 14:33:41 ok, in which case i can pick that up after finishing the ooni task 14:33:44 cool (: 14:33:47 sounds good. 14:33:57 * Tor Weather successor (karsten) 14:34:08 * irl had meant to invite the person to the meeting but forgot ): 14:34:09 there was a thread about somebody wanting to write a new weather. 14:34:36 I started writing a response but figured we should discuss first how we want to support this. 14:34:53 we can still invite them to the next meeting. 14:35:33 so, the question is what we're going to do here, ranging from nothing to adding their code to our codebase. 14:35:56 however, I feel like we're way too busy at the moment. 14:36:27 I'd think that we should encourage them to run this as a community project for a while. 14:36:40 i think it would be great if we could get a minimal tor weather running that allows relay operators to subscribe to alerts and uses onionoo as a data source 14:36:50 this would include not calling it "Tor Weather", nor running it on a torproject.org host. 14:36:59 if they want to write this in python, go or java then i'm happy to provide limited help 14:37:02 yes, I agree that that would be good. 14:37:24 right, I was thinking that we can give them support for anything onionoo-related, 14:37:34 but otherwise let them figure out how to best build the thing. 14:38:14 (I'm not even sure whether they asked for more support. it's a long thread, I only skimmed it.) 14:38:32 i think i'd like to respond with my ideas of what a minimal system would look like 14:38:48 but i don't think i'd want to say "and we'll host it" 14:38:58 another aspect we should discuss: should they really parse contact information, or just wait for operators to subscribe by themselves. 14:39:12 telling them your ideas sounds good, too. 14:39:14 i think that parsing contact information would not be great 14:39:27 I'm just afraid of making promises regarding taking their code or running it. 14:39:35 well, not afraid, but careful. 14:39:40 yep, i understand 14:40:00 okay, sounds like we pretty much agree on how to move forward. were you planning to respond to them anyway? 14:40:16 yes, but probably tomorrow 14:40:28 i was going to have a chunk of time where i write out the email 14:40:31 sure. the thread waits there for a couple days now, what's another day or three. 14:40:36 i have notes on this because i've thought of it before 14:40:56 sounds good! 14:41:21 I'm also happy to make changes to onionoo if that helps. 14:41:29 alright. 14:41:51 * Onionoo's document classes in metrics-lib (karsten) 14:42:12 I was wondering: did you consider pulling in onionoo.jar in metrics-bot? 14:42:37 the question is: should we really move around things between repositorites in order to reuse them? 14:42:42 is it unreasonably large? 14:42:55 oh, yes. 14:43:01 we could make one without dependencies. 14:43:13 the one right now includes all libraries. 14:43:23 perhaps it's good to do this first to make sure it's worthwhile moving it to metrics-lib later 14:43:26 which is different from, say, metrics-lib.jar which contains just the classes we provide. 14:44:01 I'm also thinking about this problem with respect to integrating exonerator into metrics-web. 14:44:21 which could be as simple as adding exonerator.jar as dependency to metrics-web. 14:44:47 (this was iwakeh's suggestion a while ago, before we discussed moving classes to metrics-lib.) 14:45:13 ah ok 14:45:27 hmm, quite a few aspects to consider here. how about we move that discussion to trac? 14:45:30 at some point we might want to have a proper onionoo client lib 14:45:44 which would move some of metrics-bot to metrics-lib 14:46:36 if you add all the options you can think of to trac then we can see how each would fit with the things we want to do 14:46:42 that sounds like a good idea 14:46:51 alright, will do that. 14:47:04 * Grant proposal (karsten) 14:47:19 I haven't heard back from isabela yet. 14:47:32 I'll ask at today's vegas meeting, but it might be that it doesn't happen. 14:47:51 in which case we could still put it on our next roadmap and try to get it funded later. 14:47:58 ok 14:48:00 and if that doesn't happen, just do it anyway. 14:48:08 there may also be some overlap here with a project at the university 14:48:39 oh? 14:48:45 i need to follow up there and see if there is something where they would do some of the work for the actual lookup services 14:48:51 and we would just do the collector integration 14:49:01 (and other metrics specific parts) 14:49:02 that sounds good to me. 14:49:22 currently the prof is in montreal but i will set up a meeting when he is back 14:49:36 okay, great! 14:50:01 I'll let you know what I hear today. 14:50:05 cool (: 14:50:17 okay, I think we ran out of items for today. 14:50:27 unless there's more? 14:50:42 none more from me 14:51:17 great! yay, short meetings. :) ttyl, and bye! 14:51:28 #endmeeting