19:59:16 #startmeeting anti-censorship checkin 19:59:16 Meeting started Thu Jan 17 19:59:16 2019 UTC. The chair is gaba. Information about MeetBot at http://wiki.debian.org/MeetBot. 19:59:16 Useful Commands: #action #agreed #help #info #idea #link #topic. 19:59:18 hi! 19:59:30 Pad for the meeting is here: https://pad.riseup.net/p/tor-censorship-2019-keep 19:59:38 at the top I added the roadmaps that we have been discussing 19:59:50 one for the whole anti-censorship (that can change over time) 19:59:52 antonela: what is a PRD? 20:00:01 product requirements document 20:00:03 and then one for each of the topics that we have more plans about: snowflake, gettor and bridgedb 20:00:14 (hi) 20:00:28 hi 20:00:29 ! 20:00:30 (hi) 20:00:35 window 13 20:00:36 antonela: ah! 20:00:41 all the times :( 20:00:43 all: please feel free to add anything to the pad that you want to get discussed here 20:00:48 im just trying to figure out what is the scope ahf 20:00:54 makes sense 20:00:59 hi 20:01:12 nice antonela: lt's add it to the discussion 20:01:19 thanks gaba 20:01:41 First, 2nd report. Kat5 is working on it. 20:01:52 cool! 20:02:22 The idea for this report is to specify the roadmap for anti-censorship and plans for it. 20:02:43 the PRD that antonela is working on could be very useful for the gettor part 20:02:54 (Though I also included enough background for external people to make sense of things. I hope.) 20:02:56 kat5: do you want to say something about this? 20:02:59 yes 20:03:27 We will need different people reviewing the section that they are doing right now work on and bringing the scope and plan for their work there. 20:03:38 Just that I have the structure in place, as well as a bunch of goals that I tool from the roadlist arma1 and gaba made. 20:03:54 I would like to fill in specific plans various teams have for completing the work. 20:04:06 What gaba said. :-) 20:04:38 any question about it? otherwise we move on to the network team 20:04:41 cool, i think i have some input about some of the issues we need to address around snowflake for availability of the service (running multiple instances of bridges, etc.) 20:05:45 yes, it would be great if you, ahf, can work with kat5 and me on the snowflake part of the report. Right now it is in the same git repo we had for report 1. 20:05:48 * arma1 is around and trying to keep up despite terrible internet connection (too terrible for loading pad e.g.). so, i can answer questions with high latency. :) 20:06:08 i'm gonna read the current content tomorrow morning 20:06:15 ok 20:06:18 moving to the network team. 20:06:19 and write an email to kat 20:06:26 ahf: great! 20:06:44 right now we have only the snowflake and its blockers for deployment in the network team roadmap. 20:06:56 yeah 20:07:02 and we are going to work on the team's roadmap while in brussels. 20:07:10 and maybe help out with the snowflake webextension? 20:07:19 i guess it's still in the maybe category 20:07:20 I would think we could add bridgedb and talk about other PTs during brussels. 20:07:25 yes 20:07:31 nickm, any thoughs about it? 20:07:59 marionette setup is still on my shoulders i think, i promised GeKo to look into that around christmas 20:08:05 cool 20:08:07 so we can have an instance u pthe browser team can test 20:08:22 gaba: about bridgedb, or what exactly? 20:08:28 hi, I'm here, sorry I'm a bit late, I was putting the kids to bed and then I forgot about it:) 20:08:42 We can take it on if need be: it's not bad code, and it should be fairly maintainable. 20:08:45 aobut including bridgedb and other PTs in the roadmap for network team 20:08:57 ok 20:09:14 "Other PTs" is kinda broad; do you mean developing new PTs, or maintaining the existing set, or something else? 20:09:38 we have a task to evaluate some of the reasonably new PT's or come up with a new design, i think? 20:09:48 all of that and defining scope and priorities of what we include in the roadmap. 20:09:52 the https one, the dns one, and the marionette one i believe 20:09:53 yes 20:10:07 the dns one might not exist, i can't remember that 20:10:22 Which one is the dns one? 20:10:57 ticket #15213 20:11:05 looks like it doesn't exist yet or the current link to it is 404 20:11:19 That's neither on the list nor in the report. 20:11:26 evaluate-or-design is stuff we can do. 20:11:52 yeah 20:11:54 yes 20:11:54 And would fit pretty nicely into our skills 20:11:57 that is what we have to do 20:12:09 mm, I didn't check tickets on PT in track. 20:12:35 arma1: you have some knowledge about this section i'm sure. you mentioned a whole list of things when we spoke about this first time 20:13:59 ok, anything else network team related? 20:14:27 what is our conclusion with the bridgedb part? :-) 20:14:35 do we have something we need to do right away, or? 20:14:44 not right away, it is a conversation for brussels 20:14:50 ack, cool 20:14:53 to include in the roadmap 20:15:23 metrics team: I spoke with karsten and irl this morning. They can collaborate with people on s19 after march. 20:15:38 there are a few tickets - bridgedb related that they can work with others 20:16:03 browser team, pili? 20:16:55 not sure if there's anyone else here from browser team but we can probably start working on the marionette integration in march or whenever ahf is done with his part :) 20:17:00 whichever comes last 20:17:14 cool 20:17:29 ok 20:17:38 i think in all of this we are missing the "evaluate" marionette part? one part is to run it and get it integrated, but i think there is a thinking/review process that we are missing maybe? 20:18:18 yes, this may also go into the brussels discussion for the roadmap, right? 20:18:24 we're also looking at including #28015 into the roadmap 20:18:24 yes 20:18:59 ok 20:19:00 or at least GeKo had indicated to me in the past that this should be one of the easiest of the s19 browser team tasks 20:19:23 we'll also need antonela for this so it depends on what else she's working on also 20:20:06 * antonela noted 20:20:19 ok. UX team 20:21:02 worked on snowflake ui for the webextension before christmas, talked with saint a bit about it 20:21:38 anything you need there? 20:21:45 a developer :) 20:22:15 developer for which part? 20:22:42 for the web extension part 20:23:37 is there a ticket for this? 20:23:40 did you talk with the people from cupcake? 20:23:54 antonela: we should take about that in brussels as well i think and try to figure out how much is needed there? also if we want to integrate with cupcake or do something else or? 20:24:01 samdney #23888 20:24:05 saint is the cupcake person 20:24:09 ahh :) 20:24:10 gaba, yes sait is the main maintainer 20:24:12 i think once the snowflake javascript code is good then doing the UI shim is more easy 20:24:13 i was thinking that... 20:24:14 yes 20:24:14 *saint 20:24:56 ok. I'm not totally clear who could be working on that part. Is this somethign the browser team can help with? 20:25:12 i tried to hunt devs at the tor browser meeting a few weeks ago 20:25:55 mmm 20:25:57 no success tho 20:26:14 ok. let's check about this after the meeting with pili. 20:26:31 if we want to do the UI for ticket #23888 we can do so pretty quickly. if we need to integrate with cupcake i think it's a bigger task 20:26:31 it's on the roadmap for feb 20:26:40 not sure who added it to the roadmap though :) 20:27:14 ok, roadmap for feb in the browser team? 20:27:18 ok 20:27:18 yup 20:27:21 * Samdney also have experiences with webextension and can maybe assists 20:27:32 no one is assigned though, so I'll bring it up at the next browser meeting 20:27:45 gettor, what is the situation with it right now? pili? 20:27:51 thanks pili! 20:28:07 I want to get it moved to torgit 20:28:09 nickm has the reviews for the 2 tickets related to gettor that we have. 20:28:36 ok 20:28:42 I was waiting on a reply from ilv about it but I hear he is super busy so i think we should try to move on without him and just keep him in the loop about what we are doing 20:28:53 antonela: related to the PRD on gettor, what do you need? 20:28:55 hiro: are you ok with this plan also? 20:29:01 yes, we need to bring sponsor work into the teams 20:29:01 yes I am ok 20:29:14 gaba: was the refactor reviewed? 20:29:28 let me check which numbers are the tickets that we have 20:29:45 do we have a ticket for the refactor review? 20:30:35 not that I know of 20:31:01 this is one: https://trac.torproject.org/projects/tor/ticket/28234 20:31:04 #28234 and #28091 20:31:06 gaba: we should sign together which user needs and user stories we are going to work with, we need to define a timeframe and we should have all the people who will be working on it in the same page about those 20:32:24 Personally I'd like to continue where ilv left 20:32:31 hiro: nobody is reviewing the refactor but we are reviewing port to python3 20:32:34 yes 20:32:54 ok, let's make a plan for what antonela is asking us to do 20:33:10 I think some of the logic behind gettor and the distribution model is pretty similar to bridgedb especially for some of the twistd task 20:33:12 antonela: let's talk outside this meeting and see how to move forward with gettor 20:33:26 gaba sure 20:33:37 ilv simplify the repo a lot so I think we can implement pretty easily what's missing 20:33:59 The issue with gettor is that we need developers and is not clear who is taking it from that perspective. It is not something to get into the network team workload. 20:33:59 woo 20:34:31 gaba do you mean the review? 20:34:36 or the development? 20:35:03 because for the development side I have just volunteered 20:35:13 ohh! 20:35:15 great hiro! 20:35:17 development side 20:35:24 (that's why I said "woo") 20:35:24 sorry I didn't understand that. 20:35:34 :)) 20:35:42 * gaba happy dance 20:35:43 ok 20:36:04 what exactly is the thing with gettor? only the python3 thing or more? 20:36:14 hiro: nice! 20:36:16 \o/ 20:36:33 samdney the code base has been stripped :D a lot of previous functionalities need to be re-implemented 20:36:38 moving on? 20:36:41 community team 20:36:51 ah ok. thanks hiro 20:38:01 not sure if there's anyone here from community, but we'll try to do this as part of the outreach work 20:38:26 our only issue is that if we're thinking about china-like scenarios we're not going to find them where we're going :) 20:38:36 * Samdney also belongs to the community team but is not really active there, at the moment. 20:38:52 hi Samdney :) 20:39:02 ok 20:39:12 I'm here too, but not part of the outreach efforts. 20:39:21 hi, pili 20:39:41 true... now I'm embarrassed :) hi kat5 and ggus 20:39:49 We already talked about team dependencies and we are organizing this with each team. 20:39:49 np! 20:40:16 i talked briefly this week with a contact from asia, but i really don't know how much work this would demand from community team (see alison's reply to the thread) 20:41:34 I guess we can try to make a push with the localization angle also 20:41:45 for example, https://trac.torproject.org/projects/tor/ticket/28526 many NGOs that we're in touch in global south, it's the first time or very new contact with tor project. asking them to host bridges is like some steps ahead. 20:42:23 I think kat5 had mentioned she could run with the documentation part at some point also 20:42:30 yes, I guess we have different types of organizations in different levels of possible involvemnt with Tor 20:42:59 Yep. It's on my list. I don't know how we do the "get some doing it so we can tell how good the doc is" part. 20:43:38 i can introduce people from the NGOs to somebody on our side 20:43:44 people who want to run the bridges 20:43:53 and don't know what to do or why it'll work or how to figure out if it's working 20:43:56 ggus, we can talk with technical orgs in colombia to review the doc and have their insights 20:44:01 we'd need to add it to the list of activities to run with people if we see that they may be interested 20:44:08 yup, what antonela said also :) 20:44:10 Question is who on our side has time to take this on. 20:44:53 yep. and if the answer is nobody, then, how come there is nobody at tor whose job is to interact with the NGOs who work towards freedom? :) 20:44:57 I guess it depends if we're doing it onsite or remotely 20:45:05 in many countries that we visited they didn't even have a solid relay community (like >10) 20:45:20 yes, there are specific people who can do it but not the majority of course 20:45:21 kat5: well, we have a relay advocate person. 20:46:01 ggus: would these people benefit in other ways from having access to fast bridges in the countries? do they pay less for national traffic and so on? 20:46:06 if I could speak with Phoul about adding this to the relay advocacy roadmap for 2019 20:46:27 no if... don't know where that came from :D 20:47:04 we want ngos to run bridges or to distribute them? 20:47:13 this isn't really about relay advocacy. hm 20:47:36 and what means run bridges (run in their own infra or hosted aorund the world?) 20:47:38 antonela: ngos want to help their own people use tor from inside china. they can't rely on random people running bridges because all of those bridges get blocked. so they want to run their own. and have their own people use their bridges. 20:48:05 ggus: they could all be on digital ocean as far as i know 20:48:21 the model of going to colombia and getting people in colombia to run bridges in colombia is not what this is 20:48:30 ah. 20:48:40 ngos inside china sharing bridges for china users? 20:48:47 is it also a good idea to ask NGOs to run bridges without helping them at each step of the way and then later providing support for it? (I apologize if I missed something related to this) 20:48:49 that sounds a bit scary :-S 20:48:57 lantern runs a pile of obfs4 bridges on digital ocean and hetzner, and all of their smarts are on detecting which bridges got blocked and rotating them to a new address 20:48:57 is it legal for a chinese person to run a bridge? 20:49:14 sukhe: we're not asking them. they want to do it. 20:49:27 sukhe: so yes, helping them and supporting every step is whta this is 20:49:32 arma1: ok 20:49:38 sukhe: I think the idea of this ticket is that we'd work with them to provide that support. And improve our docs/recs with their feedback. 20:49:43 samdney: yes, but, irrelevant, since the bridges are probably run somewhere else by somebody not in china 20:49:48 kat5: thanks, I missed the context 20:49:54 okey, so this ngo need technical help on setting up the bridges? 20:49:58 * gaba remembers everybody that this channel is logged. 20:50:02 reminds* 20:50:13 (Thanks gaba) 20:50:22 everybody who is still confused, go read #28526 20:50:50 ok. Do we have anything else for this meeting? questions, comments? 20:51:42 cricri 20:51:55 people are reading the ticket :) 20:52:01 si 20:52:01 hehe :) 20:52:03 ok 20:53:09 ye 20:53:15 * ahf have nothing else 20:53:40 nor I 20:53:40 ok, maybe this conversation is one to have between gus, pili, antonela and then we check about it next meeting. 20:53:47 question regarding (1) for that ticket. Would it help to have scripts or a way to automate running bridges on certain infra like digital ocean ? 20:53:54 are we gonna do these check-ins weekly? 20:53:56 hiro: yes! 20:53:58 same day, same time? 20:53:58 like a pre made image? 20:54:11 yes, checkin weekly 20:54:14 i think so. but the right person to say this is Phoul, i guess. 20:54:15 buenos 20:54:15 or there would be different things to consider based on the country? 20:54:16 hiro: (1) ? 20:54:28 (1) Document for NGOs how to easily run a few private obfs4 bridges. I've seen some guides floating around but nothing both simple and obviously official. 20:54:31 hiro: yes, a container? :) 20:54:33 ahhh 20:54:39 from https://trac.torproject.org/projects/tor/ticket/28526 20:54:53 antonela: in digital ocean they have droplets, one click install thing 20:54:54 hiro: one option is the container image. another option is that we make "apt-get install tor obfs4proxy" do nearly all of what you need 20:54:57 yes, thre are a few tickets about bridgedb deployment 20:55:01 ggus ye 20:55:07 there is one for debianize bridgedb 20:55:13 As a person that runs an NGO that operates obfs4 bridges on Digital Ocean, we would love a pre-made image! Currently we have terraform scripts, but an image is much more convenient. 20:55:16 this is not bridgedb 20:55:20 A droplet would be ideal. 20:55:24 ahh, sorry 20:55:42 obfs4 bridges 20:55:45 * Phoul is reading backchat 20:55:46 i got confused 20:56:14 ok. then we are done for today. thanks everybody! 20:56:19 hi Phoul, might not actually be something for you after all... unless you're interested! :) 20:56:25 isn't DO the company that recently started billing for traffic? 20:56:29 #endmeeting