16:01:19 #startmeeting OONI gathering 2016-10-24 16:01:19 Meeting started Mon Oct 24 16:01:19 2016 UTC. The chair is hellais. Information about MeetBot at http://wiki.debian.org/MeetBot. 16:01:19 Useful Commands: #action #agreed #help #info #idea #link #topic. 16:01:21 ahoy 16:03:04 hellos 16:03:11 hi 16:04:01 hi 16:04:17 Hi! 16:05:36 ok so let's get started with this 16:05:58 #topic Talk about potentially calling NetProbe simply “ooniprobe” prior to release 16:06:33 so I was thinking a bit about this and also in light of discussions had with people at the OTF summit I thought I would bring this up now before we do the first "official" public release 16:06:50 I get the impression OONI is in a way at this point a fairly strong "brand" 16:07:05 in the sense that people recognize is and have a general understanding of what it does 16:07:30 our original reasoning behind not using ooniprobe in the name of the android/ios version of it was that we thought it could lead to people getting into trouble for using ooniprobe 16:08:04 but I think that in light of the fact that there is an informed consent procedure that talks about OONI everywhere and in light of the fact that it's fairly trivial for somebody to link back netprobe to ooniprobe 16:08:21 it doesn't really make sense to confuse people more with another name of a product to remember 16:08:34 since in the end it's the same thing, but just for mobile devices 16:08:37 thoughts? 16:09:30 99% agree. 1% don't. that net explain what the app does maybe in this case would be useful to explain it better in the description 16:10:58 Reminds me of Firefox and Chrome being the Web Browsers... And Internet Explorer having meaningful readable name :-) 16:11:34 lol 16:11:35 I mean, NetProbe may be another traceroute/ping/wi-fi channels tool as well 16:12:42 I think that if we were to be of the "meaningful name" camp we would have picked a name that is actually more meaningful than "NetProbe". I don't think probe really tells much to most non native english speakers and the net in it could imply sort of anything 16:13:31 I vote for netprobe being renamed to ooniprobe 16:13:50 I agree that ooniprobe is a brand, and having another name for its mobile version is too confusing 16:13:53 Don't forget that an app name can be longer up to 255 (maybe Apple is less) 16:14:03 In the end I think that if we come out with an app that is called ooniprobe we will end up getting more users from the ones that already know ooniprobe, rather than getting users that don't know ooniprobe and just stumble upon netprobe and think it's useful to them due to it's name 16:14:04 also, we may reuse nice octopus with the torch for the icon in case of "ooni" brand reuse :) 16:14:10 also, as pointed out by hellais, the documentation on netprobe talks about OONI anyway 16:14:20 There are a lot of apps which uses the form "name - short description" 16:15:08 nuke_: yeah, we can use that extra text wisely to add some more context to the app name and be even more explicit 16:15:39 ooniprobe - discover your network 16:15:48 that's `nmap` 16:15:51 When editing the informed consent documentation for NetProbe, I found it quite "disturbing" that we're saying: NetProbe runs ooniprobe tests which are hosted by OONI, which is hosted by TPI :p 16:16:13 ooniprobe - detect signs of internet censorship (maybe too explicit and could get people in trouble) 16:16:14 too much indirection? 16:16:19 Check NetProbe tests descriptions here --- ah! they're ooniprobe tests :p 16:16:39 sbs_: ya 16:16:57 I love `network interference` working as it may be non-censorship. ISP Ad injection is not censorship, but it annoys a lot. 16:17:03 *wording 16:17:16 but it's probably too technical 16:17:29 darkk: I wonder exactly that 16:18:43 anyways let's not get too much into the weeds of this. I think we can defer to some other time what we put after $SOFTWARE_NAME 16:18:55 +1 16:19:09 I guess in a similar way that orbot is not named tor we could also use a different app name if one is more applicable. 16:19:22 but if we are all in agreement I would say we at least aim to do s/netprobe/ooniprobe/ pre-release 16:19:23 I will prefer though to use ooniprobe 16:19:49 anadahz: I think orbot is named that way mainly due to trademark issues 16:20:01 you *could* run debian in a chroot with the "lil debi" app on fdroid. then run ooni from there. 16:20:09 hellais: I actually like netprobe but have not read the beginning of the discussion yet 16:20:25 sbs_: ah yeah right, let me give you a paste of it. 16:20:44 So should I rename in ooniprobe (all lowercase)? 16:20:48 hellais: I am home actually just now 16:21:08 will read the log in 3..2..1.. 16:21:15 sbs_: https://paste.debian.net/889344/ 16:21:18 ah 16:21:43 hellais: stories of delayed trains 😁 16:21:48 yulax: yeah that is actually a good idea. I haven't tried installing ooniprobe with lil deb on android tbh. I can imagine the scapy depedencies creating some problems. 16:22:21 BTW, may netprobe be confused with some other netprobe? Anything else like https://resources.itrsgroup.com/Netprobe/general/netprobe_user_guide.html 16:23:06 Sorry. I've been speaking about ping/traceroute... I have not seen https://play.google.com/store/apps/details?id=georgios.chapsalis.netprobe&hl=en :) 16:23:08 hellais: depndencies are not a problem since it is a debian image, essentially its ovn install 16:23:15 darkk: haha good point 16:24:09 whoa, another RPi-based netprobe http://www.netprobes.info/ 16:25:20 lol 16:25:27 * sbs at home (hello!) 16:26:25 darkk: wow lol 16:28:32 for the sake of time I would say we move onto the next topic and maybe we can address a final the decision towards the end of the meeting after everybody has caught up to speed on this. 16:28:44 #topic Discuss this weeks' priorities 16:28:55 for the records, I've read the whole discussion on the rename now and I understand the rationale of it 16:29:13 the rename is probably a move that reduces the energy 16:29:53 sbs: great! So it seems like we have consensus on this. rename it is! 16:30:38 I guess from the dev front the priorities for this week are mainly around getting some skeleton of a airflow based pipeline going and some sort of API for the measurement interface. 16:31:02 with darkk we discussed some of these. I don't know how much of this should be discussed in here. 16:31:39 does somebody not have clear what their priorities should be and would like to discuss them here publicly? 16:32:34 * darkk thinks that the pad should have authorship marks next time 16:33:13 darkk: yeah +1 on authors next to items 16:34:11 I suspect anadahz of being the author of that line due to metadata I have. (highlights anadahz) 16:34:53 darkk: which? 16:34:56 darkk: I was the author of the 2nd item - I added it since an email was sent today asking what this weeks' priorities were 16:35:19 darkk: and thought that this could be an opportunity for any clarification needed 16:36:55 anadahz: Sorry, I thought of email with `Date: Mon, 24 Oct 2016 05:28:00 +0000` and have not imagined that level of indirection 16:37:31 anyhow, if no one wants to discuss this weeks' priorities now/here, we could proceed to the next item of the agenda 16:39:24 I guess we can do this off-the-record later perhaps 16:39:28 #topic Coordination of EG report publishing -- darkk@ 16:39:47 is there any clarity regarding coordination of the report release between OONI, TorProject & Anonymous Egyptian Media (I'm unsure if I can name it, probably agrabeli has the details) ? 16:41:29 darkk: I don't think we'll be pursuing a media strategy from the Tor Project, if that's what you mean 16:41:49 darkk: We'll be co-publishing with the Egyptian media org on the same day 16:42:05 Speaking of TorProject I meant publishing @ blogs.tpo 16:42:55 agrabeli: will you be coordinating the emission? Should me or anadahz help you? How? 16:43:32 darkk: I was recently introduced to a media org (not sure though if this is the same media org that would be co-publishing), but I haven't heard back from them yet 16:43:42 agrabeli: I think it makes sense to coordinate with blogs.tpo since the report mentions the Tor blockage in Oct. 16:44:09 darkk: It would be great if you and/or anadahz could help ensure that it's cross-posted on the same day, though we can all coordinate on this internally (I have relevant contacts) 16:44:23 anadahz: Are we sure that Tor was blocked? 16:44:43 metrics data confirm users' reports 16:44:53 I am a bit skeptical of whether or not we have evidence of tor blocking 16:45:18 anadahz: I've been editing the report, and it looks like we are mostly relying on metrics data. It might have been the case that tor usage was decreased in the cases that relays were dependent on digitalocean that was throttled 16:45:25 IMHO, simple network outage will not raise the number of bridge users 16:45:26 I think what is more likely happening is that a bunch of tor relays (including a lot of guards) are running on digitalocean due to their free unmetered bandwidth and therefore blocking digital ocean leads to a drop in tor users 16:46:02 what could have happenned is that DA was blocked in egypt (or throttled or whatever) a bunch of users noticed it and switched to using bridges. 16:46:03 DigitalOcean case was sort-of closed on 1st of September and IMHO it's different one 16:46:24 hellais: agrabeli the tor blockage didn't happened on the time range wthn DigitalOcean connections were throttled. 16:46:58 I don't think it's enough to look at tor metrics data (that also just show a drop in users) to imply that blocking is for sure happening. 16:48:14 To my knowledge we have no strong evidence of tor blocking happening a part from the drop in users from metrics. 16:48:31 we can probably try to rule out DigitalOcean hypothesis using OONI data, but I'm unsure if we have explicit measurements regarding tor being blocked 16:49:20 I think it's fine to mention this in the blog post as something unusual that has happenned, but without making any strong claim of it being a case of blocking of tor, because we don't reallt know 16:49:39 If we are in doubt we should not include the tor blockage. 16:49:52 hellais: agree 16:51:07 another question re the egypt report: It looks like we only have traffic samples showing HTTPS throttling. In this case, we drop all mention to SSH and OpenVPN throttling, right? 16:55:21 yes if we don't have PCAPs or other evidence of SSH, OpenVPN throttling we should not make a strong mention of it. 16:56:23 darkk: I recall that we had some PCAPs, do you think that are not relevant/showing OpenVPN throttling? 16:56:37 OpenVPN is probably OK to drop as there were known measurements mistakes (measurement "protocol" worth amending) 16:57:16 anadahz: I'll check, but there were some VPN-over-VPN mistakes made and I'm unsure if I'll be able to sort those things out. 16:58:06 I'll double check existing pcaps after the meeting, but I don't remember anything besides plaintext OpenVPN logs & syslog citations 16:58:13 darkk: I would say we don't invest any more time into this report. 16:58:29 I think what we have is good for a ooni blog post and we should wrap this up and publish it there 17:01:21 * darkk multitasked to the archive for a couple of minutes 17:01:34 yep we have only syslogs citations regarding OpenVPN 17:01:55 we have a lot of stuff on our plate now as we are approaching the end of the month and we should focus on those. When we will have more time available we can do a more in depth analysis of egypt (and maybe when we also have more probes deployed there and an on the ground partner) 17:03:55 hellais: darkk agreed let's not spend more time on this 17:04:12 anadahz: darkk : so to answer your initial questions: I'm editing the report and will be done with the edits tonight. We can do a pull request tomorrow and aim to publish by the end of the week. We can coordinate with Egyptians for the specific publication date, once we have agreed upon that internally. 17:04:15 lack of good SSH data is another missing thing. Do we have a checklist for alike stuff? If we don't, it makes some sense to write it down and discuss while in Berlin. 17:05:35 I mean, yeah, all these measurements should eventually land into the ooniprobe itself, so the code will be the checklist. 17:05:35 agrabeli: I suggest we do a pull request of the report some hours before the publication in the website. 17:05:47 anadahz: yep 17:07:43 hellais: do you consider valuable mentioning to blogs.tpo team that we've looked at the metrics data a bit and let them decide if it is worth publishing there? Some people at #tor were actually asking about tor being block in the Egypt on the 2nd of October 17:08:33 darkk: what more can we provide with tor people that is not already documented in the trac ticket and can be infered from looking at tor metrics data? 17:09:08 good question. yeah, nothing :) 17:09:46 heh 17:10:02 well if we get some better Tor tests we may be able to collect some good data about this in the future. 17:10:07 well, there is even no trac ticket :) 17:10:54 yeah my bad I was confusing it with this ticket: https://trac.torproject.org/projects/tor/ticket/20348 17:12:02 (If anything a good thing to do would be to summarise these findings similarly to the above ticket on trac) 17:12:12 do we have anything else to discuss before we adjourn? 17:15:14 I shall take that as a no? :^) 17:15:34 ok well thank you all for attending. Have fun! 17:15:37 #endmeeting