14:29:22 <karsten> #startmeeting metrics team
14:29:22 <MeetBot> Meeting started Thu Feb 15 14:29:22 2018 UTC.  The chair is karsten. Information about MeetBot at http://wiki.debian.org/MeetBot.
14:29:22 <MeetBot> Useful Commands: #action #agreed #help #info #idea #link #topic.
14:29:56 <karsten> https://storm.torproject.org/shared/Oh4g0hNenh65QZNWRsIe5zxpB3e0axASeSgo5hKOp2A <- agenda pad
14:30:40 <karsten> are you still adding topics to the agenda pad?
14:30:45 <irl> i'm done
14:30:48 <iwakeh> nope
14:30:53 <karsten> okay.
14:30:58 <karsten> * quick announcement from irl
14:31:19 <irl> antonela produced some new designs for the "badges" that accompany tweets/toots from metrics-bot
14:31:32 <irl> these have now been implemented and deployed and they look considerably better than my old designs
14:31:42 <irl> i've put some example links on the pad
14:32:01 <karsten> cool!
14:32:19 <karsten> looks indeed pretty good.
14:32:26 <karsten> I made a note to include this in the monthly report, too.
14:32:26 <antonela> (~‾▿‾)~ㅤ
14:32:30 <karsten> hi antonela!
14:32:36 <iwakeh> neat
14:32:38 <antonela> hey! just lurking around :)
14:32:55 <karsten> while you're here.... :)
14:32:59 <karsten> * Bootstrap 4.0 (karsten)
14:33:03 <irl> oh dear
14:33:06 <karsten> hehe
14:33:15 <karsten> so, irl and I briefly talked about an issue in bootstrap 3.3.7.
14:33:32 <karsten> and irl's suggestion was to edit the minimized js and replace a function with the 3.3.5 version.
14:33:44 <karsten> my question was: should we look into upgrading to 4.x?
14:33:53 <karsten> which is something that irl didn't want to do alone.
14:34:04 <karsten> is this something where you could help, antonela?
14:34:19 <irl> i think this would mean that we use the styleguide bootstrap fork to redo metrics.tpo's design
14:34:31 <karsten> the styleguide uses 4.x?
14:34:37 <irl> i believe so
14:34:38 <antonela> yes, the styleguide uses 4.x
14:34:48 <antonela> https://styleguide.torproject.org/getting-started/
14:35:00 <karsten> so, I guess at some point we'll have to upgrade to 4.x anyway.
14:35:12 <karsten> questions are: who does it? when do we do it.
14:35:13 <karsten> ?
14:35:25 <iwakeh> web designer?
14:35:29 <karsten> we could also ask our web designer who did the design last time.
14:35:44 <irl> i'm happy to do it with the help of hiro/antonela, but in the second half of the year, not just now
14:35:57 <antonela> after the new website, i can help on metrics i think
14:35:59 <antonela> yes :)
14:36:21 <karsten> okay.
14:37:15 <karsten> made some notes on the pad.
14:37:15 <irl> we'll get to see then also how the other portals have done integration
14:37:25 <hiro> the 4 used at the moment is also in alpha I think
14:37:26 <karsten> and until then we just hack 3.3.7?
14:37:46 <irl> if it helps to reassure you, the hacked 3.3.7 is what's been running for relay search since the theme change
14:37:54 <hiro> also you can just import the boostrap.css from the styleguide and you should have everything there already
14:38:16 <karsten> hiro: I have no idea how many things would break.
14:38:44 <karsten> okay, great. sounds like a plan.
14:38:57 <irl> (:
14:38:58 <karsten> moving on?
14:39:22 <karsten> moving on.
14:39:23 <karsten> * webstats memory issues (karsten), please confirm comment:6 in 25161 (iwakeh)
14:40:04 <iwakeh> I can prepare
14:40:08 <karsten> done.
14:40:16 <iwakeh> the shell scripts for the import.
14:40:21 <iwakeh> ok :-)
14:40:27 <karsten> I'll get new hardware in the first half of next week.
14:40:47 <iwakeh> what memory max limit?
14:40:55 <karsten> I thought about 64GB.
14:41:02 <karsten> (I'll still have to buy the ram.)
14:41:07 <iwakeh> ok, just run the import with that.
14:41:13 <iwakeh> ;-)
14:41:13 <karsten> everything?
14:41:15 <karsten> heh
14:41:24 <karsten> I can certainly try.
14:41:27 <iwakeh> well, I can try and check ...
14:41:39 <karsten> feel free to wait for that.
14:41:52 <iwakeh> the slicing idea only applies
14:41:54 <karsten> that == me trying it out next week.
14:42:03 <iwakeh> to very limited 8 to 16G.
14:42:22 <karsten> ok.
14:42:32 <karsten> sounds fine, I'll run it and tell you whether that works.
14:42:39 <iwakeh> I do a small test beforehand.
14:42:51 <karsten> ok.
14:42:54 <iwakeh> to make sure I didn't underestimate the needed RAM.
14:43:18 <karsten> sounds good. moving on?
14:43:21 <iwakeh> fine
14:43:26 <karsten> * more feedback on Sponsor 13 deliverable 2 (karsten)
14:43:39 <karsten> half of the month is over, and I haven't done anything for that sponsor yet.
14:43:49 <karsten> I'd like to continue working on the document next week.
14:44:04 <karsten> iwakeh: can you take another look at the comments and move the discussions there forward?
14:44:14 <iwakeh> sure!
14:44:19 <karsten> thank you!
14:44:30 <karsten> and do you have a google account?
14:44:31 <iwakeh> (I didn't know it was waiting for my comment)
14:44:52 <iwakeh> errg
14:44:52 <karsten> no worries. I had it on my list to tell you last week, and then you weren't here and I forgot.
14:45:09 <karsten> or, can I let you edit some other way?
14:45:15 <karsten> without letting the world edit, too?
14:45:24 <iwakeh> hhmm
14:45:29 <karsten> I can find out.
14:45:38 * iwakeh thinks about this question and mails later.
14:46:06 <karsten> okay!
14:46:15 <karsten> * feedback for #24229: Provide BGP Data Collection on Tor Metrics (iwakeh);
14:46:16 <iwakeh> I'd rather discuss on ticket than in
14:46:27 <iwakeh> these tiny comment boxes of the Gdoc.
14:46:32 <karsten> ah, hmm.
14:46:42 <iwakeh> we find a solution there.
14:46:48 <karsten> yes.
14:47:01 <iwakeh> Regarding BGP data
14:47:30 <iwakeh> I suggested a structure which can be commented on on the ticket
14:47:36 <iwakeh> the question here:
14:47:54 <iwakeh> Do we plan to update the data?
14:48:04 <iwakeh> in the near future?
14:48:15 <iwakeh> at all?
14:48:16 <karsten> hmm, fine question.
14:48:33 <iwakeh> if we do, I'd think about the steps necessary.
14:48:50 <karsten> I think we're mainly providing a place for researchers to serve their interesting tor-related data.
14:49:10 <karsten> so, if they want to update that data, we should make that possible.
14:49:20 <karsten> which shouldn't require much work, though.
14:49:21 <iwakeh> ok, so we just leave it at the scope of archiving third party data.
14:49:41 <iwakeh> yes, the steps are outlined on ticket.
14:49:55 <karsten> I'd say, let's start simple.
14:50:09 <iwakeh> it is very simple.
14:50:14 <karsten> sounds good.
14:50:25 <karsten> I'll take a look at the ticket, too.
14:51:05 <karsten> shall we discuss this more on the ticket then?
14:51:19 <iwakeh> yes, just wanted to raise attention.
14:51:25 <karsten> ok!
14:51:29 <karsten> * Make a plan for assigning points to tickets, including what points<>time factor we're going to use, if at all.
14:51:49 * iwakeh added the 'if...'
14:51:49 * karsten noticed that irl started using points.
14:51:55 <irl> i had a go at this
14:52:04 <irl> for the metrics-bot and relay search tickets
14:52:19 <irl> i used the network team points to time conversion system
14:52:51 <irl> i found that just thinking about how long something would take meant i had to think about how i would do it and i tidied up a couple of missing tickets or missing things that had to be done first, so setting parent tickets
14:52:59 <karsten> how about we all experiment with this, if we feel like it, but we don't make it mandatory yet?
14:53:11 <iwakeh> fine. I thought about this after noticing irl's point-adding, but
14:53:28 <iwakeh> it felt quite distracting and I didn't continue.
14:53:40 <iwakeh> I'll give it another try.
14:53:47 <karsten> yes, I'll do the same.
14:53:57 <karsten> regarding the conversion system,
14:53:58 <iwakeh> Maybe, for things that will wait longer before being worked on.
14:54:06 <karsten> can we agree on doing the same as the network team does?
14:54:16 <irl> https://trac.torproject.org/projects/tor/wiki/org/teams/NetworkTeam/ReleaseGuidelines#TeamCapacity
14:54:23 <iwakeh> Where was the conversion formular again?
14:54:31 <iwakeh> ah, thanks!
14:55:15 <karsten> I think it makes sense for compatibility reasons.
14:55:17 <iwakeh> the wheel exists, let's use it.
14:55:25 <irl> indeed
14:55:44 <karsten> alright.
14:55:55 <iwakeh> the priority field
14:56:09 <karsten> ah, yes.
14:56:26 <karsten> let's use it better.
14:56:31 <iwakeh> There are quite some tickets having 'high', but were not touched for weeks/months.
14:56:32 <karsten> so far we use high, medium, and low.
14:56:40 <karsten> true.
14:57:00 <iwakeh> i.e., let's also adjust when topics loose priority.
14:57:07 <iwakeh> or gain.
14:57:08 <karsten> agreed.
14:57:10 <irl> a weekly or bi-weekly review would catch this
14:57:17 <iwakeh> true.
14:57:35 <karsten> true. shall we append that to today's agenda, if we have time?
14:58:00 <irl> it doesn't have to be a team thing, just go through all the tickets and work out which ones are the most important ones
14:58:04 <iwakeh> maybe, limited to only looking at 'high' this time?
14:58:06 <irl> then accept them and work on them
14:58:31 <irl> but yes, also things that need discussion could then be discussed at the meeting
14:58:38 <irl> but you'd know what they are
14:58:45 <iwakeh> we could adjust
14:58:49 <karsten> okay, let's do a quick round at the end, just for high tickets.
14:58:53 <irl> ok
14:58:56 <iwakeh> after the meeting 'offline' this time,and
14:59:10 <iwakeh> plan reviewing the prios next time?
14:59:16 <karsten> sure, or that.
15:00:09 <karsten> okay,
15:00:10 <karsten> * Including Relay Search in metrics-web (#23518)
15:00:26 <karsten> I see there are tickets waiting for review?
15:00:41 <irl> yes, including one that actually includes relay search into metrics-web
15:00:56 <irl> but it will need a couple of changes before actually merging, which i've described in the comment
15:01:04 <karsten> ah, okay.
15:01:18 <karsten> so, I can review things later today, possibly tomorrow.
15:01:23 <irl> like the submodule points to my personal repo, which can change once the relay search changes are merged
15:01:28 <irl> ok cool
15:01:55 <irl> once you've reviewed i can put it all into one branch to merge in one go
15:02:15 <irl> it's currently two branches because the bootstrap js was still being discussed
15:02:46 <karsten> I can merge that one after this meeting, and you can rebase.
15:02:52 <irl> ah cool (:
15:02:55 <irl> that works too
15:03:28 <karsten> what's the part about freeze?
15:03:44 <karsten> is there a specific timing?
15:03:55 <irl> ah yes, so i need to update urls to make it work in metrics-web and this will break the deployment on the static mirrors
15:04:15 <irl> so ideally we don't want to change anything on the static mirrors once we make these changes
15:04:30 <irl> which i think is fine, as we would shortly after put in a redirect once we've confirmed it all working
15:04:37 <karsten> should we plan to do this on monday, not friday then?
15:04:56 <irl> i mean, the static mirrors would continue to work serving relay search
15:05:09 <irl> but if we made changes they would conflict with changes made for integration with metrics-web
15:05:30 <karsten> okay, let's see how far we get before the weekend approaches.
15:05:32 <irl> this is more a warning not to run ./update on staticiforme than anything else
15:05:48 <karsten> you mean run it for atlas? I don't do that.
15:05:54 <irl> yep
15:05:57 <karsten> okay.
15:06:11 <karsten> cool, looking forward to this. :)
15:06:15 <irl> (:
15:06:22 <karsten> moving on?
15:06:24 <irl> ok
15:06:29 <karsten> * Publishing bridge contact information: how to make a decision (karsten)
15:06:46 <iwakeh> where there any comments?
15:06:52 <karsten> yes, some.
15:07:12 <karsten> and I got some more from the research group that worked on identifying bridges last year or so.
15:07:24 <iwakeh> interesting.
15:07:26 <karsten> (not the one that leaked 2k bridge addresses as a result.)
15:07:40 <karsten> it's not a clear case. we might do it or leave it.
15:07:45 <karsten> I'm not sure how to decide.
15:08:01 <karsten> we could discuss this in the team, we could ask tor-internal@, ...
15:08:04 <iwakeh> The 'privacy board'?
15:08:09 <karsten> aha!
15:08:12 <irl> can we do a proposal to clarify some wording in a specification maybe?
15:08:40 <karsten> the change is pretty small.
15:09:06 <karsten> and we do have a specification.
15:09:06 <iwakeh> #20983
15:09:29 <iwakeh> https://metrics.torproject.org/bridge-descriptors.html
15:09:38 <karsten> but yes, I think the privacy board would indeed be a good place to ask.
15:09:39 <iwakeh> would need to be adjusted.
15:10:01 <irl> https://gitweb.torproject.org/torspec.git/tree/bridgedb-spec.txt#n55
15:10:24 <irl> so the bridge descriptors we get are different to the bridgedb-spec ones?
15:10:41 <iwakeh> https://trac.torproject.org/projects/tor/ticket/20983#comment:6
15:10:42 <karsten> no, but we sanitize them.
15:11:14 <irl> ok, so we should do a proposal to update the spec for this and make it clear that contact lines can be made public
15:11:19 <irl> and then it's a formal decision process
15:11:31 <karsten> well,
15:11:36 <karsten> we did this 4 years ago.
15:11:43 <karsten> making it clear that contact lines can be made public.
15:11:49 <irl> but it's not in the spec at all
15:11:56 <irl> there's no mention that bridge descriptors even have contact lines
15:11:57 <karsten> let me find it..
15:12:00 <iwakeh> the comment https://trac.torproject.org/projects/tor/ticket/20983#comment:6 replies to irl's link-question.
15:12:09 <karsten> that's an implementation spec how bridgedb works.
15:12:16 <karsten> it just doesn't do anything with contact lines.
15:12:17 <irl> aaaaaah
15:12:42 <irl> right ok, so we do already say these are public
15:12:42 <karsten> #9854
15:13:05 <karsten> that's where we clarified in the default torrc that contact lines may be published and indexed by search engines.
15:13:11 <karsten> we just didn't do it yet.
15:13:39 <karsten> so, how about I reach out to the privacy board, cc'ing you two, and ask if we can proceed or not?
15:13:49 <iwakeh> yep, sounds good.
15:13:50 <irl> i think that sounds good
15:14:11 <irl> especially for group identification
15:14:18 <irl> which wasn't mentioned in the earlier ticket
15:14:44 <karsten> the tor-relays@ thread has an updated list of possible advantages/disadvantages.
15:14:53 <karsten> I'll include that, too.
15:15:09 <iwakeh> fine.
15:15:14 <karsten> to be honest, I'm okay if the final decision is that we're not doing it.
15:15:27 <karsten> I just want to have a final decision and not go back to this over and over.
15:15:35 <irl> yep, makes sense
15:15:40 <iwakeh> true.
15:15:47 <karsten> okay.
15:15:49 <irl> we should also include then the option of a hashed contact line being included
15:15:57 <karsten> aha!
15:15:58 <irl> to enable group detection without giving away the contact line
15:16:02 <karsten> good idea.
15:16:17 <irl> otherwise we'll do the whole thing and someone will come back with this suggestion
15:16:18 <karsten> want to read the thread again and see if you have more thoughts?
15:16:23 <karsten> heh, yes.
15:16:35 <karsten> before I reach out to privacy board folks?
15:16:38 <irl> shall we make an email up on a pad before sending it?
15:16:42 <karsten> sure!
15:16:43 <irl> and then send it tomorrow
15:16:54 <irl> (:
15:16:57 <iwakeh> good idea
15:17:05 <karsten> yep!
15:17:44 <karsten> last on the list:  * Assigned Trac tickets (karsten)
15:17:50 <karsten> https://trac.torproject.org/projects/tor/query?owner=karsten&owner=iwakeh&owner=irl&owner=metrics-team&status=accepted&status=assigned&status=merge_ready&status=needs_information&status=needs_review&status=needs_revision&status=new&status=reopened&group=owner&max=300&col=id&col=summary&col=status&col=type&col=priority&col=milestone&col=component&col=changetime&desc=1&order=changetime
15:18:30 <karsten> we should have another one for open reviews.
15:18:52 <karsten> some of my tickets are in need of review, as are some of yours.
15:19:09 <iwakeh> true, I started adding my name to the
15:19:26 <iwakeh> reviewer field and even assign reviews to myself.
15:19:42 <karsten> sounds useful.
15:20:01 <iwakeh> the list looks doable.
15:20:01 <karsten> I'll take a look at ticket in needs_review, and also assign reviews to myself.
15:20:08 <karsten> and come up with a ticket query for that.
15:20:21 <karsten> (probably a 2 min thing, but don't want to leave you waiting here.)
15:20:47 <karsten> anything else for today?
15:21:00 <irl> nothing more from me
15:21:18 <iwakeh> all set.
15:21:23 <karsten> awesome!
15:21:35 <karsten> thank you, and talk to you next week! bye, bye.
15:21:38 <iwakeh> bye, bye!
15:21:39 <irl> bye!
15:21:47 <karsten> #endmeeting