20:00:30 #startmeeting anti-censorship weekly checkin 2019/04/04 20:00:30 Meeting started Thu Apr 4 20:00:30 2019 UTC. The chair is ahf. Information about MeetBot at http://wiki.debian.org/MeetBot. 20:00:30 Useful Commands: #action #agreed #help #info #idea #link #topic. 20:00:31 o/ 20:00:40 hello hello! 20:01:05 our pad is at https://pad.riseup.net/p/tor-censorship-2019-keep 20:01:24 i was thinking we do it today the usual way and then next time phw gets to run the session? :-) 20:01:34 sounds good! 20:01:42 :) 20:01:51 awesome 20:01:57 hey! 20:02:04 hey cohosh 20:02:29 looks like the pad is entirely updated, very nice 20:02:47 let's start by going over the roadmap 20:03:06 we have a new one that gaba put together, which is less entangled into the network team one 20:03:14 you can look at the 'in progress' column and see if you are working on those things 20:03:20 so we should probably be sure to check if there is something we have forgotten or that is missing or something 20:03:25 nice, thanks gaba! 20:03:43 yes please. Or if there is antyhing that still should be in the network team roadmap 20:03:55 Is that the top link? 20:04:24 yeah, top link under "Discussion" 20:04:27 on line 15 and then line 37 again 20:04:38 * sysrqb lurks 20:05:10 Got it. Thanks. 20:05:31 so this should be bridgedb + snowflake, right? 20:05:40 and PT interface tickets we keep under the network team roadmap? 20:05:41 kat5: it should be the same as we had before 20:06:38 So this is everything except GetTor? 20:06:47 yes 20:07:04 ahf:: i'm looking at the tickets you mentioned 20:07:16 mentioned where? 20:07:23 "and PT interface tickets we keep under the network team roadmap?" 20:07:36 i think they are related to dormant mode and PT's iirc 20:07:43 but the network team roadmap wont open for me 20:07:56 I left #29267 and #29283 20:08:02 yes! those two 20:08:03 very good 20:08:33 can we spot anything that is off with the new roadmap or does it reflect what is going on? 20:09:05 looks good to me 20:09:24 * ahf too 20:09:39 I left a few more in the icebox that were for may and didnt have time to move but i will 20:09:54 Did we remove the Marionette ones? I'm not seeing them. 20:10:01 yeah, that is cool. it's nice to have things together and not in multiple docs 20:10:17 kat5: i think we are ignoring marionette a bit right now after we found out that John H is no longer with redjack 20:10:22 kat5: i think we decided to take marionette off the roadmap for now 20:10:27 ah yeah ^^ 20:10:31 Okay. Thanks. 20:11:18 okay, let's go to the discussion points then 20:11:37 hm, gaba, did you write the bridgedb one? do you want to run that one? 20:11:56 yes 20:12:11 there are quite a lot of tickets that were assigned to dgoulet during brussels meetings 20:12:30 this is not so much a discussion but an announcement :) 20:12:41 dgoulet may not go back to those tickets after he comes back 20:13:10 ah, right, because of the v3 onion services work that needs to happen when he is back? 20:13:46 i can start assigning myself to some of these tickets, and hopefully slowly move towards maintaining bridgedb eventually. 20:13:49 yes 20:14:04 phw: that would be awesome 20:14:11 cool! 20:14:12 I'm sure he will be available and may be able to do some but not most of them 20:14:20 phw: sysrqb can be od great help there, he knows a lot about that stack 20:14:25 s/od/of/ 20:14:26 available for questions I meant... 20:14:41 :) 20:14:44 ahf: right, there should still be plenty of help 20:14:59 yep 20:15:06 cool! 20:15:24 okay, the next item we have is from nickm who brought this up during the network team meeting monday 20:15:34 (hi) 20:15:37 about handing over ticket #29269 to the anti-censorship team 20:15:38 hi nickm 20:15:59 maybe you want to talk a bit about this ticket 20:16:40 as i understand it , the goal of this ticket is to look at the information that bridges export in their routerinfo and extrainfo descriptors, and see whether it meets the needs that metrics and anticensorship have for such information. 20:17:16 one way to analyze what info is available would be to go over the stats code in tor and make sure it all is still turned on (or can be turned on) for bridges 20:17:34 a simpler way, though, would be to just ask metrics what info is available for bridges... 20:18:01 ... and then to ask anticensorship and metrics if they want more 20:18:02 or different 20:18:04 or other 20:18:11 (end of description) 20:18:35 and once we know a bit more concrete if we are missing something or if something should be changed we create core tor/tor tickets for the tasks? 20:18:40 right 20:18:43 so we takeover this one as a meta-ticket 20:18:51 we might want to defer that till after privcount is farther along, even 20:19:10 we have this one marked as sponsor 19 right now, so maybe we want to change that 20:19:11 As a sample, here's what raw info bridges have in their extra-info descriptors: https://collector.torproject.org/recent/bridge-descriptors/extra-infos/2019-04-04-20-09-01-extra-infos 20:19:34 And here's an example of taking that data to get client country counts for Snowflake: https://trac.torproject.org/projects/tor/ticket/29734#comment:4 20:20:46 this seem okay to everybody? 20:20:50 yeah 20:20:55 yes 20:20:56 we should put this on the roadmap 20:21:06 it is on the roadmap in the backlog list 20:21:10 ah 20:21:19 coolio 20:21:32 My turn? 20:21:44 yep :-) 20:21:57 The report started out being based on arma's list of priorities from December. Then I started adding tickets tagged as sponsor19. Then we refocused on the 4 roadmaps and that's where we are today. 20:22:18 Some of the stuff from arma's list includes community outreach, ux under bad network conditions, and better understanding censorship of Tor. 20:22:35 phw and I were discussing removing that content from the version we submit to the sponsor, but including it in another version we publish for the general community. So first, we'd like to know if folks think that's a good idea. 20:23:20 what is the idea behind splitting it up? like having two variants of it? 20:23:31 the sponsor don't care about ux/outreach stuff? 20:23:37 a bit of context: i found report 2 very useful to get up to speed on anti-censorship work and i think it can be a great resource for prospective volunteers. these volunteers may care about more than just the funder. 20:23:46 Well, we don't have concrete plans to do it at this point. 20:24:18 I think the idea was that the main things for the sponsor should be the stuff we're doing near-term. 20:24:22 ah yeah, and the funding period is ending soon 20:24:27 Yep. 20:24:54 but we could just leave it for funders and public, right? 20:24:59 makes sense 20:25:00 this sounds good to me, i also found report #2 very useful 20:25:17 :-) 20:26:06 gaba: we could keep content such as "community outreach" for both funder and public, yes, but it may be slightly out-of-scope for the funder. 20:26:47 the funder should know that community outreach is important for all the other work :P 20:27:02 I think we didn't so much tell the funder what would be in any of these reports. And that was on purpose. 20:27:41 I'm good with either, so will wait for someone to tell me which to do. 20:27:50 what does arma1 think of this plan? 20:27:53 if he's around 20:28:10 i think he knows this funder quite well and maybe have some hints on how to navigate it 20:28:24 yes, sounds good 20:28:29 kat5: sounds great. in this case i'm leaning towards not making our lives harder, and having a single report that keeps the content as it is. 20:28:38 +1 20:28:45 phw: Okay! 20:29:10 also, cheers to everyone who contributed to it. it's very clear and well structured 20:29:45 Cool. That 20:29:48 's all from me. 20:30:02 cheers to kat5 :) 20:30:03 cool 20:30:14 \o/ 20:30:17 okay, then we have the "Help with" sections of the pad 20:30:36 phw has a good one: Help me understand what you're up to and how I can help. 20:31:03 phw: what are you interested in here? should we do a round saying what we are up to or is it more a generel thing of getting to know what is going? 20:31:19 i'm still in the process of figuring out what everyone's up to, and what i should focus my attention on. 20:32:08 on a very high level, i think we need to get better at understanding censorship of tor, to make good decisions going forward 20:32:31 +1, very much 20:32:50 phw: yeah it seems some of the reachability stuff you have started on is good and important to do soon 20:33:01 as a concrete example, we don't seem to have a good understanding of bridgedb metrics. what's the distribution of email addresses that ask for bridges? without answers to some of these questions it's difficult to focus on the most pressing things. 20:34:12 yeah 20:34:26 that's just a quick note on my first impression after four days of work :) 20:34:54 i think all of that is very important with the one of the team goals that is to handle things that happens suddenly 20:35:02 yep, makes sense me 20:35:40 right now we have focused on things on the backlog that we need to do: the snowflake/obfs4/bridgedb work, but we haven't touched much of the "getting a better understanding with data collection" yet 20:35:47 i guess this is also quite related to nick's ticket 20:36:03 there are some roadmapped items about stats collection for bridgedb 20:36:30 i think snowflake is super important and we should keep pushing in this direction. 20:36:45 we've been doing mostly snowflake work recently on the anti-censorship team 20:36:59 other than dgoulet's work on some pressing bridgedb tickets 20:37:20 but with me we have an extra 40h of work and i want to use some of it to better understand our systems and users -- for now with respect to bridgedb. 20:37:22 and hiro on gettor 20:37:34 ah yes ^^ 20:37:39 hi roya here for short time. 20:37:43 cool, that sounds good! 20:37:46 phw: i think that is very great. i guess that also relates a bit with info we can collect from our friends from ooni 20:37:53 hello roya 20:37:56 hi roya 20:37:57 hi roya o/ 20:38:26 related to this ticket: "https://trac.torproject.org/projects/tor/ticket/29277" happy to help 20:38:28 fyi, roya is a professor at umich who may be able to help us with bridge scanning using her project, censored planet. 20:38:44 ah, very cool 20:39:45 roya: basically, what we could provide as input is a set of bridges, i.e., ip:port tuples. 20:40:21 for each tuple, it would be great to learn in what country the ip:port seems unreachable on the network level 20:40:40 are there specific countries you are interested 20:40:43 And OONI already has the set of bridges for their measurements: https://github.com/OpenObservatory/ooni-resources/blob/master/bridge_reachability/tor-bridges-ip-port.csv 20:41:12 how often do you need the tests to run, daily? 20:42:05 roya: ideally all countries :) 20:42:46 we can get Augur (spooky scan ++ ) run soon. Obviously we can from all, but there are 50,000 vantagepoints to choose 20:43:03 daily measurements for the bridges in dcf1's list would be a great start. 20:43:12 how does this work? you have nodes in the respective countries and you can for bridges from there? 20:43:27 s/can/scan/ 20:43:35 so far we exhaust vantagepoints in not free countries (42 based on Freedom House report) and then choose samples from others 20:44:10 ahf: https://www.youtube.com/watch?v=uelf0yDMA1E 20:44:24 ahf: way cooler than that, see https://censorbib.nymity.ch/#Pearce2017a 20:44:48 I use advanced side channels to bounce packets off of specific routers 20:45:05 thanks, gonna watch/read that! 20:45:49 I thought bridges IPs are private, aren't they 20:45:51 ahhh, i remember that talk. cool 20:46:08 roya: most are, but some are hard-coded in tor browser. these are effectively public. 20:46:14 There are private ones, but the ones in the list I linked are the Tor Browser default bridges; i.e., not really private. 20:46:45 That list also includes directory authorities and some other stuff, plus I'm sure many of the bridges are offline, FYI. 20:46:46 for measuring them, should we worry about bulk tests coming out of UMich machine 20:48:03 i can ask all details question later, sorry don't want to hijack the meeting, But just know we are ready to support this 20:48:32 roya: ok, shall we use the trac ticket to coordinate this? 20:48:59 Sure 20:49:16 o 20:49:32 i'll add a summary of our discussion to the ticket 20:50:56 cool 20:51:08 i think we went over the items kat5 had on the help with list? 20:51:12 * Samdney is also here and watching. 20:51:16 Yep. 20:51:53 then we need to create a ticket to get phw access to snowflake.git, i think i can do that using the template sysrqb used the other day for the sponsor 19 repo 20:52:18 gonna do that post-meeting 20:52:27 phw: Is there already a ticket to get you push access to the report repo? 20:52:42 i think it has been granted already, i think i saw that on the internal channel 20:52:48 kat5: i already have access to sponsor19.git and pushed my minor changes 20:52:53 Excellent. Thanks. 20:53:00 thanks ahf 20:53:12 cohosh would like help with #25688 20:53:18 help with review* 20:53:30 I'm going to look at that probably today. 20:53:35 dcf1: thanks! 20:53:36 cool! 20:53:59 catalyst: i also got the wekan link that didn't work - the link that works is the top one on the pad that we used earlier in the meeting 20:54:02 i was a bit confused by that too :-S 20:54:13 i think it was storm behaving like storm does 20:54:20 Sorry. I had some problems with Wekan 20:54:43 i don't think we have any more help items on the pad. is there anything else we should talk about while we are here? 20:54:48 One more topic to bring for next meeting is how we are going to do reviews in the team. Right now reviews are being managed by asn and dgoulet at the network team. 20:54:51 then now is open floor :-) 20:54:57 But it can be discussed next meeting. 20:55:05 ahf: yes the link in the pad works now :) 20:55:08 (5 minutes for an hour of meeting) 20:55:11 gaba: cool! 20:55:20 anything else? :-) 20:56:03 Hi... Is it OK to chip in here? I'm Karl from Internews, working on Pluggable Transports 20:56:12 hello karl! 20:56:14 yes it is 20:56:23 hi o/ 20:56:28 you were at our last meeting in mexico, right? 20:56:37 Yes, that's right 20:56:43 cool, welcome to our weekly meeting 20:56:54 Hi again to those I've met / spoken to, and thanks to cohosh for inviting me 20:57:38 I'm really keen to make sure we're all joined up with PT work, so will hopefully be able to join in more. For now, I'd like to mention that we have some new funding available for PT development 20:58:07 cool 20:58:18 The form is linked from https://pluggabletransports.info 20:58:21 we can pass the info over to our grand writers 20:58:46 thanks kk-int 20:58:53 kk-int: i have a question for you that i think you might be able to help me with since you know the PT world very well 20:59:08 kk-int: do you know what has happened with redjack and marionette? i tried to email John H some days ago and the email bounced 20:59:22 1 min for the hour of meeting :P 20:59:32 :-D 20:59:51 I don't know the story behind it, but I believe he is no longer there and to my knowledge they have not continued with marionette work 21:00:03 right, that was a bit our conclusion too :-( 21:00:10 but we hadn't managed to confirm it 21:00:14 ahf: did you try contacting scott coull or tom shrimpton? they are co-authors of the paper. 21:00:30 (i think kevin dyer, the main author, is with google now) 21:00:42 phw: no, i did not. only john since he showed up and was very into it suddenly. he joined the tor dev meeting in rome in early 2018 too 21:01:07 gotcha 21:01:22 BTW we've created a session tomorrow at IFF for circumvention tech - https://platform.internetfreedomfestival.org/en/IFF2019/public/schedule/custom/1517 - anyone not there can join in via #traffic-obf 21:01:23 idk how much the other paper authors are involved in the development work, but we could check 21:01:42 oh, there is a traffic-obf irc channel 21:01:43 kk-int: cool, thanks! 21:01:54 oh, maybe it's not an irc channel, but a hashtag? 21:02:26 ok, gonna close the meeting now as we are the hour, but we can continue chatting in here 21:02:31 #endmeeting