14:16:48 #startmeeting Measurement Team meeting, take 2 14:16:48 Meeting started Wed Aug 12 14:16:48 2015 UTC. The chair is karsten. Information about MeetBot at http://wiki.debian.org/MeetBot. 14:16:48 Useful Commands: #action #agreed #help #info #idea #link #topic. 14:16:56 1. add Sean to the Tor Tech Report 14:17:04 yep 14:17:07 2. you can add me to a +, or leave me off if there's already too many authors. 14:17:14 either is fine. 14:17:19 will add you. 14:17:27 https://people.torproject.org/~karsten/volatile/measurement-roadmap.pdf 14:17:32 ^ latest version 14:17:43 at some point I'll clean up the author block. 14:18:11 Thanks 14:18:17 how can I contribute to the report? 14:18:19 so, let me tell you what I think could be next steps there. 14:18:38 SeanSaito: depends on the contribution. if you want to make minor edits, send me a patch to.... 14:19:00 https://gitweb.torproject.org/user/karsten/tech-reports.git/tree/2015/measurement-roadmap/measurement-roadmap.tex?h=measurement-roadmap 14:19:23 if you want to make major edits, obtain a lock by sending me an email, edit edit edit, and then send me a patch. 14:19:41 so, my plan for this roadmap: 14:19:58 1. get more feedback from team members within the next couple of days or week. 14:20:34 2. ask people on tor-relays@, tor-dev@, tor-talk@, dir-auth@, etc. what features they want us to provide. 14:20:57 3. ask fundraising folks how we should set priorities between user groups or even specific features. 14:21:28 4. determine a set of projects for the next X months, think about finding volunteers and/or funding for them. 14:21:49 what do you think about this plan? 14:23:19 (if you prefer to read the current draft after this meeting and send feedback via email, that works for me, too.) 14:24:58 For 2, how do we plan to talk to users from Group 1 (the general public) in figuring out what features they would like to see? 14:25:31 fine question. we might ask the people in tor who periodically talk to the general public. 14:25:44 like those giving talks and doing outreach. 14:26:18 Right. And do we plan to have meetings with the Network Team as well in the future? 14:27:17 ah, no, but we can ask them to give us a general idea what things they are interested in that doesn't change too quickly. 14:27:40 that being said, you can always join the network team meetings as well if you're interested. 14:28:47 Interested. When are they held? 14:28:52 to be clear, I don't think we'll have a perfect picture of what users want from us. 14:29:12 but asking some users will ensure that it will be a better picture than something that we make up ourselves. 14:30:00 * karsten tries to find out 14:30:35 https://lists.torproject.org/pipermail/tor-news/2015-August/000108.html 14:30:47 there are some upcoming events at the bottom. 14:31:14 I don't think there's one network team meeting. but the little-t meetings are pretty close. 14:31:28 and the Pluggable transports/bridges meeting belongs to Network Team, too. 14:31:52 not sure whether these meetings are weekly or bi-weekly. 14:32:38 okay, should we play a round of 1-1-1 task exchange while we're here? 14:32:47 1-1-1 task exchange: you get 1 minute to describe a task that would take somebody else roughly 1 hour and that they will do for you within 1 week (review a document, write some analysis code, fix a small bug, etc.; better come prepared to get the most out of this; give 1, take 1) 14:33:02 virgil, SeanSaito, qbi: ^ 14:33:32 I could start. 14:33:45 sure, please do! 14:34:15 * karsten will collect tasks here: https://etherpad.mozilla.org/2U0TTCtYwu 14:36:23 SeanSaito: do you mean start by describing a task or by picking the first task? :) 14:36:39 So I am currently implementing the gamefication features, the core of which is the badges/points allocated to each family based on their stats. The badges currently include bandwidth, consensus weight, exit bandwidth, size of family, geo diversity of relays, liberal exit policies, # of guard/exit relays and existence of contact info. 14:36:54 Sorry, describing the task (more than a minute has passed I guess) 14:37:06 no, go ahead. 14:37:18 it's the first time we're playing this game. 14:37:45 Put simply, I would like to have more ideas/categories for awarding badges/points. 14:37:52 * qbi doesn't really understand it. 14:38:14 so, you provide a list of what you have, and the task is to add/edit that list? 14:38:22 But I'm too distracted (phone calls, people coming into the office etc.) 14:38:51 qbi: SeanSaito is looking for ideas for characteristics that make a relay awesome (or less awesome). 14:39:01 Yes. Well, this is something I just thought of right now. 14:39:05 feel free to explain in more detail if you want. qbi would be a great person to think about this. (fresh brain!) 14:39:44 One idea I discussed with Virgil was to add a badge that indicates whether a relay is eligible for a Tor t-shirt. 14:39:46 (sorry for being late to the meeting. nothing to report from my side, however.) 14:40:11 hi phw! you can still participate in the 1-1-1 task exchange. 14:40:17 So it would be nice to add some fun badges like that too, in addition to the more important ones. 14:40:28 okay, SeanSaito, added to the pad. 14:40:47 should we collect more tasks first before we find people for them? 14:40:58 virgil: thx 14:41:19 SeanSaito: feel free to tweak the text on the pad. 14:41:41 who's next? 14:41:55 (feel free to pre-write your task description and paste it here.) 14:42:02 I have a 1-hour fixable thing. I'd like to know if anyone is using Onionoo's extended_family field. 14:42:09 no wait--effective_family 14:42:27 I know that nusenu was using it. 14:42:34 okay that answers it. 14:42:40 done. 14:42:43 heh 14:42:53 yes, let's leave it in and consider removing it in a few months from now. 14:43:18 * karsten adds a task: 14:43:21 Investigate two display bugs in ExoneraTor's permanent link (#16588, #16589); some experience with Bootstrap would probably help. 14:43:54 this is really something for a web designer. 14:43:57 which I'm not. 14:43:58 the MyFamily field is collected from the individual micro-descriptors, correct? 14:44:12 server descriptors, not microdescriptors, but yes. 14:44:48 more tasks? 14:45:07 another 1hour thing would be add some client-side sanity checks to those. E.g., if the fingerprint is longer than 40 characters, truncate it. 14:45:39 or, let's turn that into "find out in which places it's possible/feasible to specify a relay with more than just its fingerprints" 14:45:56 k. 14:46:03 that's all I got. 14:46:50 I have one more for people who like writing scripts: 14:46:57 Automate the process of checking MaxMind's website for new GeoIP databases and updating tor's src/config/geoip* files (e.g., #16687). 14:47:11 the current steps I'm doing are here: https://people.torproject.org/~karsten/volatile/README.geoip.local 14:47:46 qbi, phw: do you have anything you want other folks to do for you? 14:48:21 karsten: no. 14:48:29 ok. 14:48:50 no 14:48:56 ok. 14:48:59 https://etherpad.mozilla.org/2U0TTCtYwu 14:49:07 who wants to take what? :) 14:49:30 I think I could take virgil's task. 14:49:57 I'm not sure if it's so useful to take SeanSaito's task, because I already did a braindump on that a while back. 14:50:12 having somebody else do that would probably be more useful. 14:50:33 * karsten takes virgil's task. 14:50:58 let me clarify the 1-hour bit. 14:51:12 the idea is that you plan to take an hours to produce something useful. 14:51:27 if the task takes you 1:10h, that's fine, too. 14:51:45 but if you realize that you won't produce anything good after 15 minutes, stop doing it. 14:52:03 I can have a look into the display bugs, but not sure if I'm webdesigner enough to finbd a fix 14:52:06 i'm happy to think about SeanSaito's task. 14:52:09 maybe tell the task owner why you thought you could do it. 14:52:12 Karsten, I could do your MaxMind task 14:52:24 but I don't know where do update src/config/geoip 14:52:25 sounds great! 14:52:49 qbi: no worries if you don't find out. 14:52:55 good 14:53:05 phw: Thanks! 14:53:28 phw: do you have enough context, or do you need more info from SeanSaito? 14:54:23 * Sebastian arrives very late 14:54:24 or, more generally, if somebody lacks context, just ask via email. 14:54:29 i think i have enough context, otherwise i'll drop sean an email. 14:54:30 hi Sebastian! 14:54:33 if anyone needs anything from me or I can help you, let me know :) 14:54:35 phw: great! 14:54:48 Sebastian: any tasks you want somebody else to do similar to these? 14:54:51 https://etherpad.mozilla.org/2U0TTCtYwu 14:55:53 anything else we should discuss today? 14:56:18 figure out all places where threads are started inside Tor 14:56:32 and under what conditions 14:56:35 maybe? :) 14:57:00 sure, why not. 14:58:06 * karsten figures that "how do I notice when a thread is started" almost disqualifies himself for this task.. 14:58:36 does anybody else feel more qualified? 14:59:28 how about we leave it on the list for next week? 14:59:56 sounds great to me 15:00:01 perfect! 15:00:15 okay, I think that's it for this week. let's talk again in a week from now! 15:00:20 karsten: I've decided I'm going to fix the definition of median 15:00:20 thanks for coming, everyone! 15:00:26 in the tor sourcecode 15:00:36 virgil: is it broken? 15:00:53 or just not specified as clearly as it should be? 15:01:01 for ordered list {a,b,c,d}, it returns b instead of (b+c)/2. 15:01:24 yes. maybe that's for a reason (which I don't know). 15:01:40 I look forward to hearing this reason when my patch is rejected. 15:01:41 like, using value (b+c)/2 would break for some reason, whereas any of a, b, c, d would be fine. 15:01:45 you cannot do that 15:01:51 without breaking Tor's voting 15:02:21 Tor's specification requires low median for a bunch of directory stuff 15:02:41 sounds like a fine post-meeting discussion. again, thanks for coming, everyone! :) 15:02:44 #endmeeting