15:00:26 <asn> #startmeeting SponsorR
15:00:26 <MeetBot> Meeting started Tue Apr 21 15:00:26 2015 UTC.  The chair is asn. Information about MeetBot at http://wiki.debian.org/MeetBot.
15:00:26 <MeetBot> Useful Commands: #action #agreed #help #info #idea #link #topic.
15:00:32 <dgoulet> hello meeting
15:00:34 <asn> greets
15:00:37 * nickm in attendence, distracted, cheerful
15:00:46 <isabela> hehe
15:00:48 <nickm> (and bad at spelling)
15:00:54 * syverson likewise I guess
15:00:55 <asn> :)
15:01:03 <asn> ok i see ohmygodel i see syverson i see dgoulet
15:01:10 <asn> maybe not karsten, but he sent an email about his potential absense
15:01:28 * teor just ended up in the SponsorR meeting by hanging around too long
15:01:31 <asn> so let's start with status reports
15:01:45 <asn> during past week, I looked at the graphs in #15513
15:02:01 <asn> i ended up comentoring two students in SoP
15:02:17 <asn> i got plenty of feedback on the direct onion services proposal
15:02:46 <asn> i opened #15743 and #15714. the latter has split up further.
15:03:03 <kernelcorn> and I was accepted for SoP under Yawning and asn, I proposed the Onion Name System
15:03:04 <asn> i'm still udnecided about #13667.
15:03:13 <asn> and that's that from me I think.
15:03:27 <asn> i have a few ideas about further stats, but they are not very good.
15:03:31 <asn> who wants next?
15:03:32 <ohmygodel_> is kernelcorn jesse victors ?
15:03:36 <asn> ohmygodel_: correct
15:03:43 <ohmygodel_> ok
15:03:50 * ohmygodel_ can go next
15:03:51 <asn> kernelcorn: ohmygodel_ is aaron johnson.
15:03:52 <asn> ohmygodel_: please go
15:04:02 <ohmygodel_> i went to the sponsor r meeting last week
15:04:17 <ohmygodel_> mostly david and rob did actual work
15:04:36 <ohmygodel_> but i am working with SRI to collect timing information for their crawler
15:04:43 <ohmygodel_> to see what kind of performance they are seeing
15:04:46 <asn> i saw that. waht's the end goal?
15:04:47 <asn> ah i see
15:05:04 <ohmygodel_> our whole justification for improving HS performance
15:05:18 <ohmygodel_> depends on such stats
15:05:27 <asn> yes.
15:05:30 <ohmygodel_> thats also why Rob made OnionPerf
15:05:33 <asn> this might be handy client-side stats indeed
15:05:38 <asn> are they using one-hop circs?
15:05:47 <asn> for crawling?
15:05:54 <ohmygodel_> no
15:05:58 <asn> (shouldn't they?)
15:06:11 <ohmygodel_> maybe ?
15:06:15 <asn> ack
15:06:20 <ohmygodel_> maybe they do care about anonymity though
15:06:31 <asn> since they are using full circs, then the client-side stats might be representative
15:06:39 <ohmygodel_> they keep getting kicked off various sketchy sites for being too obviously botlike
15:07:10 <asn> no surprises here
15:07:17 <asn> fair enough
15:07:18 <ohmygodel_> also i worked on the updated technical report about HS stats obfuscation techniques
15:07:28 <ohmygodel_> yes i havent forgotten about that
15:07:29 <teor> and no robots.txt equivalent for .onion sites
15:07:48 <ohmygodel_> actually its basically done, just some minor thing to add before sending it to you all
15:07:53 <asn> ohmygodel_: ack
15:08:02 <ohmygodel_> its about 12 pages now
15:08:19 <ohmygodel_> im not sure how to get it pulled into tor’s official git repo
15:08:35 <asn> pulled into? pushed into?
15:08:36 <ohmygodel> but maybe somebody can advise me on that later when you get a chance to look at it
15:08:51 <ohmygodel> pulled, pushed, shoved, slid, whatever
15:08:57 <asn> i don't think there is an official git repo for tech reports.
15:09:02 <asn> or maybe there is and i just don't know
15:09:22 <asn> i think karsten can figure this out.
15:09:26 <ohmygodel> well then maybe i can just pull request to https://git.torproject.org/user/karsten/tech-reports.git
15:09:30 <asn> yes
15:09:31 <ohmygodel> … if i know how to do that
15:09:39 <ohmygodel> ok FIN
15:09:40 <asn> you can't really pull request like on github
15:09:43 <asn> you would have to open a trac ticket
15:09:45 <asn> or send karsten an email
15:09:55 <asn> ohmygodel: ack thanks!
15:09:56 <asn> next?\
15:10:26 * syverson can go I guess.
15:10:30 <asn> syverson: please
15:10:35 <kernelcorn> clone to Github, then make pull request
15:10:45 <kernelcorn> or make a diff or .patch
15:10:48 <syverson> Attended Sponsor R QPR. Worked with SRI and others on challenges a bit. Taught some people there more about how Tor works and why. Began a response to the "Why is Tor working with DARPA question" that was much improved by Roger and issued by him. Fought with ohmygodel about naming and design issues for [successor-to-hidden] services. Attended the PM departure celebration and talked to the new PM about our goals.
15:10:53 <syverson> Done.
15:11:26 <asn> great thx. you also answered the popularity thread, but i haven't had time to look at it yet.
15:11:30 <asn> next?
15:11:34 <dgoulet> o/
15:11:39 <asn> dgoulet: pleas
15:11:40 <asn> e
15:12:07 <syverson> Right something I forgot. Probably forgot other things, but others please go.
15:13:05 <dgoulet> QPR like you all know so some PR and helping other people understanding Tor, did some practical work on #15714 and #14917
15:13:21 <dgoulet> FYI, #15714 is now split into 3 tickets you can find them as child ticket
15:13:34 <dgoulet> and also I'm now wiser on SMC thanks to ohmygodel  :)
15:13:38 * dgoulet done
15:13:41 <asn> thx
15:13:48 * nickm did very little for R, but tried to follow the various discussions.  I wrote an implementation of PCSA for my approx_counting branch. Maybe others will find it useful.
15:13:50 <nickm> Done.
15:13:55 <asn> PCSA?
15:14:04 <nickm> Probabilistic Counting with Stochastic Averages
15:14:05 <asn> the hash counting trick?
15:14:08 <asn> ah ok
15:14:14 <nickm> the trick on top of the hash counting trick
15:14:23 <nickm> where you have multiple counters that you choose between.
15:14:23 <asn> :)
15:14:50 <asn> ok so discussion phase
15:14:55 <asn> what would you like to discuss?
15:15:04 <asn> i have a few ideas on further stats, but they need to be refined.
15:15:11 <asn> we could talk about that if you want.
15:15:11 <dgoulet> roadmap refinement?
15:15:18 <isabela> yay!
15:15:21 <asn> roadmap refinement sure
15:15:35 <ohmygodel> asn: i would like to hear about that
15:15:36 * isabela guess what topic I have (top down list :))
15:15:37 <kernelcorn> dgoulet: nice find on #14917, well done on that
15:15:41 <asn> ah https://trac.torproject.org/projects/tor/wiki/org/sponsors/SponsorR
15:15:45 <asn> because the one in here ^
15:15:48 <asn> is up to April
15:16:10 <asn> ok then it probably makes sense to refine the roadmap
15:16:26 <asn> fwiw, my exam period starts these days, so I will be low activity for the next 10 days. and after that I will be super low activity for 2 weeks.
15:16:30 <asn> just saying :)
15:16:35 <dgoulet> asn: ack
15:16:54 <kernelcorn> asn: you know, a HS monitoring system is easier if you have a DNS, then it's very easy to get lists of HSs
15:17:15 <Yawning> meep
15:17:24 <asn> ok. so let's talk abit about more stats
15:17:32 <asn> since that's going to be useful during roadmap making
15:17:43 <asn> so we did #15513, which I think was great.
15:17:55 <asn> dgoulet: did you have any HSDir-related graphs in the end?
15:18:34 <dgoulet> asn: nope unfortunately, I still plan to try to have some by the end of april and actually have the hs health running full time
15:18:42 <dgoulet> 15513 is very useful yes!
15:18:45 <asn> ok
15:18:59 <asn> so hsdir stats are coming up. that's ok.
15:19:13 <asn> after that, we are undecided on the next stats
15:19:16 <asn> here is the top down list: https://docs.google.com/spreadsheets/d/1mY8wax7FBUIAPAmDLpGcEtYCYCcsxW21KtLCntRbSk0/edit?pli=1#gid=0
15:19:27 <asn> some ideas from that list:
15:20:00 <asn> - on the stress testing part, we could play with the profiler results from #15515 etc.  and identify choke points and places to be improved
15:20:17 <asn> we have already done a bit of this, and we could present some pie charts and stuff on where the CPU is spent on a busy HS.
15:20:28 <asn> potentially useful. potentially useless too.
15:21:04 <asn> - still on #15513 , I am a bit curious on looking how the bandwidth graphs of IPs of busy hidden services look like.
15:21:11 <dgoulet> asn: hmmm we have already tickets for the chockepoint we could start working on
15:21:19 <asn> like we could check the IPs of the hidden services when they have lots of traffic, and see if we notice any bumps on their bandwidth graphs.
15:21:40 <asn> but that's not too exciting either.
15:22:02 <asn> - i'm still curious about #15714 and how we could find a better number than 16384
15:22:23 <asn> we could try to think of more statistics that could help us here, without completely leaking popularity
15:22:32 <asn> but this does not seem doable cleanly, without an SMC system.
15:22:55 <asn> another stupid suggestion would be to run those statistics on a few of our relays for a few weeks, and locally check the results
15:23:10 <asn> in that case we are not worrying about the stats beeing forever or global.
15:23:20 <asn> and it would give us a better idea on how the numbers look like.
15:23:29 <asn> but it's very cheap to run stats only on our own relays. so i don't really like this idea either.
15:23:31 <dgoulet> asn: I can see that working (for IPs)
15:23:40 <ohmygodel> asn: i think local, private measurements is a useful technique in general
15:23:54 <asn> ohmygodel: i think so too, but it's a bit rude maybe.
15:24:09 <asn> i think it's useful because without some initial meausremnets it's hard to understand if our fears are reasonable or not
15:24:57 <ohmygodel> academic tor researchers do it all the time
15:24:58 <asn> anyway, these are my ideas so far.
15:25:03 <asn> ohmygodel: true
15:25:12 <ohmygodel> (that doesnt make it not rude, though)
15:25:37 <ohmygodel> so are you going to fix that popularity leak ?
15:25:41 <asn> which one?
15:25:42 <ohmygodel> by removing the IP cell cap ?
15:25:54 <asn> i think nickm and dgoulet are workking on the 16384 one
15:26:03 <dgoulet> ohmygodel: #15745 ;)
15:26:25 <dgoulet> but we still need to decide if we want to ditch officially the balancing algorithm
15:26:48 <asn> i think we should. at least in its current form.
15:27:07 <ohmygodel> dgoulet: a randomized cap is good, no cap is better
15:27:14 <asn> the fact that it doesn't have memory, makes it rotate through a big number of IPs quickly.
15:27:17 <dgoulet> ohmygodel: not sure no cap is better...
15:27:24 <ohmygodel> for privacy anyway
15:27:37 <dgoulet> asn: we did have a ticket for ditching the algo rith? (can't find it :S)
15:27:43 * kernelcorn is now AFK
15:27:51 <asn> that's #4862
15:27:59 <dgoulet> asn: great
15:28:30 <asn> anyway, i'm not totally against running local measurements to figure out better numbers for this ticket.
15:29:20 <dgoulet> no idea how private measurement will give us a better number for 16384 tbh...
15:29:39 <asn> we would check how many intros we get in intro circs for a month
15:29:50 <asn> and see if we get any of them with more than 16384.
15:30:08 <asn> if lots of circuits see more than 16384 intros, then that number is probably too low.
15:30:27 <dgoulet> ah! wait "private measurement" means run that on our own relay?
15:30:30 <asn> yes
15:30:33 <dgoulet> (ok sorry I thought it was in private network)
15:30:59 <dgoulet> yeah I think we could definitely do that
15:31:55 <asn> ok
15:32:00 <asn> that could potentially be a next stat then
15:32:12 <asn> because the profiler output is not that exciting.
15:32:13 <asn> this might be more fun.
15:32:55 <dgoulet> asn: will you have time to make a patch for this?
15:33:04 <asn> ehhhhmm
15:33:13 <dgoulet> :)
15:33:22 <asn> maybe
15:33:25 <asn> if we decide what it needs to do
15:33:40 <dgoulet> asn: ok let me know if you can't make it, I can jump on it I guess
15:33:41 <asn> but i would prefer to do it in 3 weeks tbh
15:33:52 <asn> although it doesn't seem too hard.
15:33:57 <asn> ok
15:34:09 <asn> let's use the ticket
15:34:11 <asn> in the following days
15:34:12 <asn> to discuss this
15:34:16 <asn> and see what we should measure
15:34:19 <dgoulet> asn: yeah I think it's going to be pretty straight forward so we could do it soon-ish and run it for a while, I just don,t want to put you in trouble for your Uni ;)
15:34:24 <dgoulet> asn: ack
15:34:54 <asn> simply counting intro1s per circuit would be very straightforward. making the collection a bit more privacy-preserving would require more thought.
15:35:19 <asn> ok
15:35:23 <asn> let's use https://trac.torproject.org/projects/tor/ticket/15744 for this task.,
15:35:30 <dgoulet> hash the onion with siphash24() and count them so when we log we don't know which .onion (anyway ticket talk :)
15:35:41 <dgoulet> perfect
15:35:46 <asn> IPs don't get to learn the onion anyway
15:35:51 <asn> ok
15:36:10 <asn> let's talk on that ticket for the next week or so. if we decide on something good, maybe we can hack something up soon.
15:36:15 <asn> otherwise, I will do it when I finish exams.
15:36:20 <dgoulet> ok
15:36:43 <asn> so OK, one potential controversial stat got added.
15:36:52 <asn> maybe we can move to roadmap refinement?
15:36:53 <nickm> If this is onion counting we can do even better for privacy if we accept some uncertainty. See my approx_count branch.
15:37:26 <asn> nickm: i'm not very familiar with the approx_count technique
15:37:42 <asn> but i'm not sure if it adds something here
15:37:57 <ohmygodel> nickm: its not onion counting - its introduction cell counting at an introduction point
15:37:59 <asn> it's about not keeping IPs, when we count unique IPs right?
15:38:37 <asn> shit IP is an overloaded term in this conversation.
15:38:48 <ohmygodel> yeah there is no count unique streaming items issue here, which nickm’s approximate counting is for
15:39:11 <nickm> so it's count-INTRODUCE2-cells-per-IP ?
15:39:26 <ohmygodel> INTRODUCE1, i believe
15:39:35 <asn> yes. probably intro1.
15:39:42 <nickm> ah
15:39:43 <asn> maybe per-service.
15:39:46 <asn> or per-circuit.
15:40:03 <ohmygodel> cant be per-service (ips dont know service), right?
15:40:14 <asn> correct
15:40:27 <asn> it would be per-circuit. (and assuming that it's one circuit per service)
15:40:33 <dgoulet> hrm isn't there the service pub key in the establish_intro cell ?
15:40:59 <asn> that's a fresh key for each IP, not the normal pukey of the hidden service
15:41:21 <asn> IIRC, each HS keeps an assymetric keypair for each one of its IP.
15:41:38 <asn> the pubkeys are exposed in the descriptor.
15:41:45 <asn> this is to not leak the identity of the HS to the IPs. it's smart.
15:41:55 <kernelcorn> clever
15:41:57 <ohmygodel> asn: you could keep a running noisy histogram and add in each circuit’s count once it’s closed
15:42:18 <asn> ohmygodel: yes, that would be the verbose version.
15:42:29 <ohmygodel> you would still have a running total for an open circuit, though
15:43:06 <asn> i'm not completely against a histogram, provided that the time period is big enough. maybe a week or a month.
15:43:39 <asn> the more privacy-preserving technique would be a simple 'True' or 'False' on whether any circuits saw more than 16384 INTRO1s during the past week.
15:43:40 <ohmygodel> if you dont want the histogram, the a noisy running total count of all INTRO1s is even easier, of course
15:44:03 <asn> total count would not help us figure out whether the 16384 number is good, right?
15:44:43 <ohmygodel> asn: im not sure how paranoid you need to be here. the worry is that you get hacked at some point?
15:44:56 <asn> no
15:45:01 <syverson> Who's "you"?
15:45:05 <ohmygodel> asn: yeah, you're right.
15:45:06 <asn> just that the results should not be too revealing when we publish them.
15:45:09 <ohmygodel> “you” = asn
15:45:27 <syverson> grrr
15:45:30 <ohmygodel> oh then fuck all this shit
15:45:32 <asn> (we are planning to publish these results, right?)
15:45:46 <ohmygodel> store everything you want and publish nothing
15:45:57 <ohmygodel> just use it to inform your protocol improvements
15:46:30 <dgoulet> asn: hrm I wasn't thinking we should publish those
15:46:32 <asn> that's also a tactic. store everything for our use. and in the end publish a sanitized report.
15:47:35 <asn> anyway, let's use #15744 to discuss these techniques, ok?
15:47:56 <syverson> Hmmm. That seems a pretty serious divergence from general Tor openness principles.
15:48:05 <asn> i agree
15:48:21 <ohmygodel> i disagree
15:48:26 <asn> haha
15:48:30 <ohmygodel> paul you yourself did this
15:48:33 <syverson> Surprise!
15:48:42 <ohmygodel> in your measurement study with matt edman
15:49:05 <ohmygodel> you stored a bunch of shit during the study, and then only published very aggregate statistics afterwards
15:49:31 <syverson> Right because we realized that we were not as privacy preserving as we had thought.
15:49:50 <ohmygodel> and rob is doing this with his retransmission observations on micah’s relay
15:49:57 <ohmygodel> and ivan did this with his HSDir observations
15:50:05 <ohmygodel> and gareth owen did it with his HSDir observations
15:50:26 <ohmygodel> and RPW did it with his onion router TCP connections observations
15:50:29 <ohmygodel> etc. etc. etc.
15:50:33 <asn> it's true. but i guess we could display higher openness standards.
15:51:08 <syverson> I'm not saying we shouldn't do this at all. I'm saying to have eyes open and recognize what principles you're following. How you might be breaking them. If that's a good idea. How you should revise your principles. etc.
15:51:15 <ohmygodel> oh and roger did it when he started collecting HS stats on moria during the sponsor r kickoff
15:51:36 <ohmygodel> this is well established practice
15:51:57 <syverson> It may be that the oft-stated principle is crap. That then requires revisitng.
15:52:00 <ohmygodel> not that the issue is closed
15:52:06 <ohmygodel> just its hardly “pretty serious divergence from general Tor openness principles”
15:52:27 <asn> oft-stated?
15:52:36 <syverson> Roger says it all the time.
15:53:00 <syverson> We should only collect anything that we would be willing to publicly share.
15:53:23 <asn> ok ok
15:53:29 <asn> let me think about it for the next few days
15:53:36 <asn> and i will suggest a few different techniques in the ticket
15:53:43 <ohmygodel> interesting - i believe roger was actually collecting HSDir *fetches*, which we know of no way to safely publicly share
15:53:57 <syverson> ohmygodel: I was underscoring the problem with the principle and drawing a bright line between principle and practice, which is not good. Kinda my point.
15:54:09 * nickm 's principle is subtly different: If I would accept a hypothetical patch that prevented anybody from collecting a statistic, then the information is private. :)
15:54:19 <nickm> but this principle too is suspect
15:54:20 <ohmygodel> i dont believe in that principle, and i dont know anybody that does
15:54:27 <ohmygodel> in fact violation of it is built into Tor itself
15:54:39 <ohmygodel> because Tor collects data for bridge statistics that it would never publish
15:54:56 <syverson> I just wanted to note it. I don't think this is the time for this discussion.
15:55:20 <asn> nickm: what do you think about collecting info for solving  #15744 on our relays?
15:55:36 <asn> nickm: if we collected that info on the whole network, we would leak popularity, which we have decided not to do, supposedly.
15:55:47 <asn> nickm: do you think that collecting suhc research info on our own boxes for a few weeks is too naughty?
15:55:52 <syverson> For your values of "we".
15:55:59 <asn> tor devs
15:56:30 <asn> nickm: we would then look at the info, analyze it to derive conclusions, and publish whatever we think is appropriate to the public.
15:56:46 <syverson> I disagree with you asn that tor devs decided this except perhaps by fiat of some, but let's leave that aside for now too.
15:57:44 <nickm> let's defer the topic, I say
15:57:53 <asn> ok
15:58:00 <asn> let's talk in #15744  about this
15:58:01 <asn> over the next week
15:58:02 <asn> or so
15:58:03 <ohmygodel> (btw nickm, i was referring to syverson’s principle, not yours. your msg came out while i was writing that)
15:58:08 * karsten is around now
15:58:13 <asn> hello karsten!
15:58:18 <karsten> hi asn
15:58:20 <asn> eok
15:58:21 <asn> ok
15:58:27 <asn> so does it make sense to do roadmap refinement now?
15:58:30 <asn> or leave it for next week?
15:58:39 <syverson> That is *not* my principle. I'm just the messenger.
15:58:42 <dgoulet> asn: hrm couple of things
15:58:45 <asn> dgoulet: do say
15:59:06 <dgoulet> tor perf tickets, seems Rob took the ball on that which would differ them after April
15:59:17 <asn> ok
15:59:24 <asn> what is rob's thing btw?
15:59:28 <asn> oniionperf or something?
15:59:29 <dgoulet> OnionPerf
15:59:31 <asn> is it like torperf but better?
15:59:36 <asn> or different use case?
15:59:38 <dgoulet> I hope he updates tor-dev soon with it
15:59:41 <asn> ok
15:59:54 <dgoulet> yeah it's TorPerf 2 but with a service component also as far as I know
15:59:55 <ohmygodel> it currently fetches some files of a fixed size over a hidden service he’s running
16:00:04 <ohmygodel> btw i talked to rob yesterday
16:00:17 <ohmygodel> he doesnt want to tackle torperf 2 as described in the proposal
16:00:24 <Yawning> (offtopic, I really need to steal dgoulet for a day from sponsor R and lock him in a room with all the accumulated torsocks patches)
16:00:28 <ohmygodel> its like tedious software engineering stuff
16:00:41 <asn> there is a toprerf2 proposal?
16:00:50 <ohmygodel> e.g. split into these modules, use these packaging techniques, etc.
16:00:51 <asn> writen by someone who is not Rob?
16:01:09 <nickm> tedious software engineering stuff is my bread and butter, my rice and beans, my pasta and tomatoes.
16:01:10 <Yawning> ohmygodel: hire slave^wgradstudents?
16:01:11 <karsten> there's a tech report where we sketched out a design for a torperf 2.
16:01:11 <dgoulet> Yawning: ack, I'm *way* to far behind on that, need to talk half day to merge all and release, please continue to pressure me this week if I don't do it ;)
16:01:19 <asn> karsten: aha
16:01:20 <nickm> that said, my time is hyperlimited over the next N months.
16:01:52 <dgoulet> ohmygodel: ok so Rob just wants the quick results of it or improve the proposal or ?
16:02:07 <ohmygodel> sorry nickm, we’re not lazy, but we don’t get rewarded for spending much time on that kind of stuff
16:02:59 <asn> dgoulet: so what about torperf?
16:03:08 <asn> dgoulet: why did you mention it i mean?
16:03:28 <dgoulet> it's in April HS roadmap
16:03:35 <asn> ah
16:03:41 <dgoulet> we had kind of plan to have it on metrics soonish
16:03:42 <ohmygodel> dgoulet: if someone in tor is excited to incorporate this into metrics measurements, then i think it can go forward. if not, we’ll keep collecting the stats for our use in sponsor r only.
16:04:08 <karsten> we could start by adding a link to metrics and think about doing more.
16:04:11 <ohmygodel> but actually you should talk to rob about it
16:04:28 <karsten> a link being a page with a short description that can be found on the main metrics page.
16:04:31 <dgoulet> ohmygodel: ack
16:04:44 <karsten> ohmygodel: sounds good.
16:04:58 <dgoulet> karsten: right but we really don't want to use current TorPerf right, instead maybe use R to build v2 ?
16:05:42 <karsten> dgoulet: depends on who's going to do it.
16:05:59 <karsten> current torperf is a bunch of scripts. better not touch anything or you'll break it.
16:06:14 <dgoulet> yeah that's my concern ^ ... ENOTENOUGH resource/time I think
16:06:17 <karsten> building v2 requires some effort, but may save a lot of trouble long-term.
16:06:59 <karsten> but did we promise anything in april, which ends in about 9 days?
16:07:24 <dgoulet> karsten: not that I know of no
16:07:29 <karsten> ok good.
16:07:46 <asn> i think we are ok yes.
16:07:57 <asn> anyway, let's do roadmap refinement next week ok
16:07:58 <asn> ?
16:08:07 <ohmygodel> sounds good
16:08:18 <ohmygodel> i may stop coming to these meetings for a while
16:08:19 <asn> btw it's likely that someone might have to run this meeting during the first week of may.
16:08:34 <ohmygodel> bc im working on stuff now that doesnt really involve you all much
16:08:41 <asn> ohmygodel: ok
16:08:52 <dgoulet> ohmygodel: you will still continue on SMC a bit?
16:09:05 <ohmygodel> dgoulet: yeah, thats a summer project
16:09:06 <asn> dgoulet: karsten: you have things for next week?
16:09:09 <dgoulet> ohmygodel: great
16:09:16 <dgoulet> asn: yeah this week is booked for me so no worries
16:09:27 <asn> ok
16:09:35 <karsten> asn: I'll spend a bit of time on that new intro point graph and look at onionperf.
16:09:44 <karsten> should I be doing anything else? (haven't read backlog yet.)
16:09:59 <asn> maybe CC yourself to #15744?
16:10:07 <karsten> ok.
16:10:09 <asn> i will try to post some technqiues there this week
16:10:10 <isabela> asn: I can help with the meetings while you are away
16:10:27 <asn> isabela: great. will tell you.
16:10:32 <asn> ok thanks for the meeting folks
16:10:35 <asn> have a good day!
16:10:36 <asn> #endmeeting