16:01:32 #startmeeting 16:01:32 Meeting started Mon Oct 17 16:01:32 2016 UTC. The chair is anadahz. Information about MeetBot at http://wiki.debian.org/MeetBot. 16:01:32 Useful Commands: #action #agreed #help #info #idea #link #topic. 16:01:44 helo! 16:01:49 :) Hello everyone 16:01:54 ehlo! 16:02:02 darkk: :-) 16:02:18 hellos :) 16:03:41 OK, so now that we are all around let's have a look at the today's agenda. 16:04:11 #topic Get ooniprobe packages in stable Debian Stretch release before the freeze 16:04:35 is freeze date announces? do we know the deadline? 16:05:21 darkk: Yes "After 5th January 2017, removed packages will not be permitted to re-enter testing." (https://release.debian.org/stretch/freeze_policy.html) 16:05:48 2016-11-05: Transition freeze (general freeze of transitions) 16:06:02 anadahz: do you think we should rush to also make MK into stretch? 16:07:43 For those of you that don't know what freeze means: Debian "stops" the introduction of new packages entering in the current stable release (i.e, Stretch). 16:08:00 what's transition freeze? is it `no new packages in testing` or `now new transitions from unstable to testing unless it's a bugfix`? 16:08:19 That's why we had no ooniprobe package in Debian Jessie :( 16:08:38 sbs: I think yes it makes much sense to rush! 16:09:10 irl: are you happen to be around? 16:09:14 okay, so I need help to make the Debian package :) 16:11:18 darkk: good question though I don't really know the difference. 16:12:22 darkk: This is even more confusing: https://wiki.debian.org/Glossary#transition 16:12:31 is ooniprobe a good thing to package to `stable` for a year and a half? what's the benefit? isn't it too moving? 16:13:05 darkk: why do you say a year and a half? #IgnoranceHere 16:13:28 darkk: The benefit is that people can install ooniprobe without mixing up testing/unstable repositories. 16:14:41 sbs: I just looked at https://en.wikipedia.org/wiki/Debian_version_history#Debian_8_.28Jessie.29 16:14:43 For a time frame we had a number of users running a very old and outdated ooniprobe version since there were no new packages in stable (still is the case). 16:16:28 anadahz: do we want to have a separate repository for these users with latest ooniprobe that is useful for debian stable dependency-wise? 16:16:31 darkk: I see (sigh, reading that news reminded me that now debian ships with systemd :() 16:17:25 darkk: We used to have a separate repository but it was a PITA properly update and maintain. 16:17:37 anadahz: how many those users are? (in any case I guess it makes sense updating the ooniprobe in stable not least not to keep in stable a super old version) 16:18:48 sbs: https://qa.debian.org/popcon.php?package=ooniprobe 16:19:17 anyway, we already have ooniprobe in `testing` so there is no way back, it'll land in stable and (probably) the best thing we can do is to replace 1.6 with 2.0 :-) 16:20:08 anadahz: danke schon -- btw, I cannot find ooniprobe in stable, why? https://packages.debian.org/search?searchon=names&keywords=ooniprobe 16:20:26 darkk: agree! 16:21:20 sbs: note that the package popularity contest is not enabled by default so it is highly probabe that are/were more ooniprobe package installations in Debian. 16:21:36 anadahz: noted! 16:22:05 anadahz: does debian-based lepidopter report to popcon? 16:22:12 sbs: We didn't make it prior to Jessie freeze :( 16:22:55 darkk: no, also lepidopter is not using the debian package of ooniprobe but the pypi one. 16:23:12 yep, that's true 16:24:00 sbs: do you have a ticket for the MK Debian package? 16:24:16 is there a reason to package MK separately from ooniprobe? ... to package MK before ooniprobe migrates to MK and is ready to be packaged to debian? 16:26:27 anadahz: I have this repository from which I can create a debian package (https://github.com/measurement-kit/measurement-kit-debian) but the process is very "ghetto" 16:26:46 darkk: the second you said 16:27:52 anadahz: I see (referring to the Jessie freeze) 16:28:16 sbs: does MK shares the similar dependencies with ooniprobe? 16:29:16 darkk: I guess they will be two separate packages: MK and ooniprobe? 16:29:22 anadahz: nope, but the things it depends from are in debian for sure 16:30:07 anadahz darkk: yes, my understanding of debian packaging is that they would like the two to be separate packages 16:31:46 the second question is more interesting: does it make sense to package MK for distro if ooniprobe (it's only user that is packaged) does not use it yet? Am I missing something? 16:31:59 anadahz: MK depends from libevent, libcrypto, libssl, libz 16:34:05 darkk: I don't know... let's assume ooniprobe starts using MK in five months, at this point: a) can we update ooniprobe in stable and b) is MK in stable required in order to do step 2? 16:35:28 I know almost nothing about stable release updates in Debian. There are updates like 7.x, but I don't know if they can contain non-bugfix packages. 16:35:35 In any case we 'll *need* to have ooniprobe in Debian stable it's nice to just instruct users to install ooniprobe by apg-geting it ;) 16:35:54 anadahz: true 16:37:29 Given the fact that MK could be used by ooniprobe it makes sense to add MK as well(?). 16:38:42 Yep, but we should not forget that the price is "being stuck with some number of 2.0 probes for another year or two". See also "tor at Ubuntu" case :-) But, yeah, you're right, there is some number of users that will be OK with apt-getting the probe but will unlikely setup anything more complex to get the latest version. 16:39:17 guys, what about backports? 16:39:59 if, say, we bless ooniprobe 3.0 with MK and we cannot put this in backports because MK is not in debian stable, I'd say this sucks (for how many users, tho?) 16:41:08 darkk: Once the package is inside the stable repository we can have security or normal updates. Security updates are being applied ASAP, normal updates are following the standard Debian package release. 16:41:53 sbs: I guess we do have time to ensure MK is in stable right? 16:42:24 do rules prevent us from pushing LIBS into backports? 16:43:37 I doubt that's true. IMHO, if the package is in the backports and it requires some fresh library, the library should also be packaged for the backports and that's it. I'm not debian lawyer but I'll be surprised to hear that my reasoning is wrong. 16:43:40 anadahz: yes, we just published a stable release and we can do some more testing, perhaps bless another point release (0.3.3?) and try to have that in stable 16:45:56 darkk: makes sense 16:48:26 darkk: yeah we can always do that 16:49:48 So it seems that we have decided upon making Debian packages for MK and ooniprobe ASAP. 16:50:15 anadahz: yep 16:50:30 what else do we have on the agenda? 16:51:37 no more agenda items (https://pad.riseup.net/p/ooni-irc-pad) 16:54:21 okay, then perhaps we can adjourn, unless is there anything else to discuss? 16:54:38 In the week of 7/11 to 12/11 we are going to be meeting in real @ Onionspace in Berlin. 16:55:17 just a note: I've circulated a pad with you for creating an agenda for the upcoming OONI internal meeting & public hackathon. please add things there that you would really like to work on, so that we can add them to the agenda. :) 16:56:06 Official announcement will follow shortly.. in case any oonitarians are aeound we 'll be happy to meet you. ;) 16:56:06 I had a conversation with a couple of the google people last week about their interest in making a visualizing map in the space 16:56:09 we'll post an announcement of the OONI (public) hackathon somethime this week (hopefully) 16:56:11 i think they've talked with hellais as well 16:56:37 but having unified front for knowing what to tell them is useful for them to be doing would be good 16:56:40 willscott: yes, he had a meeting with them last week 16:57:02 (it's the group i interned with a couple years ago, so i know all of them pretty well) 16:57:13 my worry with where they are now is that they currently are thinking of building their map as google 16:57:16 i don't think that's what we want 16:57:24 i think we want them to contribute the visualization as open source to ooni 16:57:32 willscott: do you mean those people from Project Jigsaw? 16:57:35 yeah 16:58:01 Then I don't get what `map` you mean. 16:58:09 willscott: to my understanding the call with google mainly involved hellais communicating what OONI does, its goals, next steps, etc. 16:58:28 okay, i can tell you what my conversation involved: 16:59:01 they're thinking of making a visualization map of censorship with data from ooni, iclab, etc. 16:59:41 they have a bunch of ux resources, and having them contribute such a map would be nice in the sense that they would likely make something that is functional and that looks nice 17:00:09 that said, i don't think having it be a thing run by google has some problems 17:00:39 like, activists won't trust it as neutral as much (at least from china / places with big anti-google sentiment) 17:01:02 other new types of research won't be able to as easily compare their data sets against existing ones if it isn't open source 17:01:43 so, as you keep interacting with them, important to keep talking aout how you see the eventual visualization within ooni existing 17:02:13 and why it's important for that to be open source to encourage other academic and non-profit projects to overlay & compare their data sets through the same analysis and visualization system 17:02:45 willscott: yeah, I think their discussion also involved the eventual visualization, though hellais can provide you with more info on that next week when he's back 17:03:15 yup 17:03:16 willscott: agreed 17:03:26 willscott: agrabeli: are they going to be in Berlin on the meetup dates ? 17:03:35 i would not expect them to be 17:03:41 anadahz: no 17:03:54 anadahz: they're based out of new york 17:04:22 i think they're pretty early in their process, and not ready to engage at that level yet 17:04:41 they mean well, just need to hold their hand and channel that energy effectively :) 17:05:42 willscott: I think they can undoubtedly help with creating some great visuals, and I also think they can help in terms of data. That said, you're very right to point out that things should remain open source for all the reasons you outlined. 17:05:57 willscott: Thanks for letting us know! That could be really interesting/worth looking into. 17:06:25 IMHO, if the visualization is the cornerstone then it sounds like `give us your data and we'll do a nice picture and use it in our PR machine` to me. But Jigsaw seems to be a bit broader than that so I'm likely wrong :) 17:07:22 i think it's stemming from the same motivation all of the existing projects have, of wanting to have a map of what's being censored where 17:07:42 darkk: at the end of the day OONI data is publicly available for everyone to use. ;) 17:07:45 i think they also want to help the ecosystem though 17:09:08 anadahz: yep, I know that the data is open. I don't get visualization story. It looks like it's just a part of the conversation and I'm missing the fullview :) 17:10:29 I think darkk has a point 17:10:56 but I also agree with anadahz and willscott in terms of OONI data being available anyway, everyone wanting a map and so on.... 17:12:06 overall, I think we all want there to be a great visualization out there which maps censorship in the most accurate way possible -- whether that's done by OONI, Google, or whoever 17:12:29 so based on that, if Google can do a better job at creating such a visualization, so be it 17:12:57 however, there may be some conflict of interest here, in the sense that OONI's latest funding proposal is about creating precesily this thing 17:13:24 *precisely 17:15:23 agrabeli: Yep, let's discuss this further when we have more facts/the complete overview in hand. 17:15:25 I can hardly imagine useful "map of censorship". It's just a single-color globe with neutral water in the middle (sorry, read too much Chomsky :-)) On the other hand I can imagine map of network interference that is telling some story about the technology & ISPs. 17:16:04 details :p 17:18:21 in any case, I think it's too soon to make a call on this 17:18:39 absolutely. just something to be thinking about going forwards :) 17:18:40 to my understanding hellais just had an initial introductory call 17:18:49 willscott: indeed :) 17:20:36 so if there's nothing else to discuss, perhaps we can end this meeting (20 minutes overtime)? 17:22:09 agreed 17:22:31 Thanks everyone for attending! 17:22:40 #endmeeting