13:29:04 #startmeeting 13:29:04 Meeting started Wed Oct 21 13:29:04 2015 UTC. The chair is nickm. Information about MeetBot at http://wiki.debian.org/MeetBot. 13:29:04 Useful Commands: #action #agreed #help #info #idea #link #topic. 13:29:07 hello tor friends! 13:29:14 hi 13:29:24 hi 13:29:25 8 weekdays left in october! 13:29:38 I'm going a bit crazy 13:29:45 and all will be okay; 8 days is plenty 13:30:32 status: I'm trying to get all the sponsor U stuff done. I did the DoS stuff at a first draft level, and am doing a second patch before I make it public. Andrea's helping with one of the proposals, and writing one of her own. I'm pretty happy with dev doc status, and worried about guard stuff.... 13:30:53 ...though isis and athena are both interested in the guard stuff so that's a good sign 13:31:00 also trying to put 0.2.7.4-rc out today. 13:31:06 first I'm trying to squash all the leaks in master 13:31:09 so many leaks 13:31:10 is it just a pile of docs that needs to be done? 13:31:31 mostly design documentation, 13:31:44 with the main coding being a prototype guard algorithm implementation 13:32:13 hi meeting! 13:32:16 (to force us to say what our best idea for a guartd algorithm is. Could be in python if we can test its behavior. 13:32:19 hi athena ! 13:32:37 and if I get all this stuff done, I'll be a happy person and find some other nice thing to do. probably more code review for folks. 13:32:42 who's next with a status? 13:34:33 would it work better if I just start picking people? 13:35:16 Yawning: how's it going with you? 13:37:04 ok, gonna go back to my eternal battle with documentation 13:37:23 and start trying to wrap my head around what needs to be done in 0.2.8 13:37:50 at least there's light at the end of the irl distraction tunnel some time 13:37:57 I'm expecting that battle to not be repeated. we're gonna have to do more pt evaluations, but the plan there is Don't Make Yawning Do 'Em. 13:38:14 I'll do em if needed 13:38:17 unless you secretly love doing pt evaluations 13:38:20 because someone has to 13:38:36 sure, but if we diversify who _can_ do them, we can use you for more stuff 13:38:40 and you can have more fun 13:38:53 I am in docs hell this month and it's not as much fun as I'd hoped 13:40:13 anything else? any docs stuff we can help you with? 13:40:55 think I'm good 13:41:01 ok. athena ? 13:41:04 I will scream in panic if I belive otehrwise 13:44:35 so for me, I'm in recovery mode of a wrist issue, but since last week I've done mostly prop#250 code (it's working, we are able to have a shared random value within the consensus), I've done few things related to HS also, I still plan to do some code review by the end of october for 027 stable 13:45:24 cool. is the proposal on tor-dev/ in tor-spec up to date? 13:46:18 nickm: the tor-spec one needs some changes but minor, it will be up to date by the time we submit the code for review 13:46:37 and asn is working on the one for the ed25519 shared random key 13:47:20 neat! 13:47:31 one helpful thing to do when you put the code for review... 13:47:36 status: we have an apparently-functional DoS filter for dirauths at last (ticket #4581); i'm still doing some unit tests, tweaking the default parameters and trying to make a proposal for sponsorU out of it 13:47:47 is to review other big patches. I am WAY behind. ( 13:47:49 :( 13:48:05 athena: cool! The proposal is the part that would be an end-of-month deliverable 13:48:15 did you have a chance to read the thing I set yesterday btw? 13:50:12 nickm: will read this afternoon 13:50:17 thanks! 13:50:31 Any feedback you can give to help me finish it up would rock 13:51:01 I know you've thought about this at least as much as I have 13:51:11 (the doc is a proposal for refactoring and hiding authorities) 13:52:45 anybody else with a status today? 13:52:54 if not, do we have anything for discussion? 13:55:11 oh meetings! i can actually attend them now because they are at a sane time! 13:55:19 because DST shifted? 13:55:28 wait no that makes no sense 13:55:31 because you moved? 13:55:32 it did? 13:55:37 because i moved 13:55:46 one more reason for everybody to leave the west coast 13:55:59 isis: hi! want to drop a quick status update for us? 13:56:37 i started working on #17261 yesterday, and i expect that tomorrow i should have something to send into #tor-dev 13:57:23 rocking! 13:57:26 i also realised that this paper: 13:57:28 https://github.com/isislovecruft/library/blob/master/cryptography%20%26%20mathematics/post-quantum%20cryptography/A%20Quantum-Safe%20Circuit-Extension%20Handshake%20for%20Tor%20(2015)%20-%20Schanck%2C%20Whyte%2C%20Zhang.pdf 13:57:49 has an implementation: https://github.com/NTRUOpenSourceProject/ntru-tor 13:58:05 and mostly for kicks i was taking a look at it 13:58:07 kinda 13:58:24 there's something wrong with it? 13:58:25 it does the handshake stuff, but not the part that's annoying which is oversized cells 13:58:34 unless they changed the code 13:58:40 ah, i see, because the NTRU keys are huge 13:58:45 also IIRC it uses the ref implementation 13:58:46 of NTRU 13:58:53 which needs changes iirc 13:59:04 yep 13:59:33 i don't think the code should be merged, obviously, but i was impressed that they actually wrote up the code and published it 14:00:26 and that is about all i've accomplished this week because i'm in the process of moving 14:00:52 I think you also looked at my aez silliness 14:01:07 ah yes, i need to finish that email 14:01:56 nickm: your proposal is far from silly, but the AEZ security proof as i mentioned is rather scary 14:02:17 the proposal actually seems fine 14:02:43 There's a funny quote in the spec about their proof 14:03:04 the "anti-easy" part? :) 14:03:19 From their Table of Properties for AEZ 14:03:32 "Proofs" : "Either: Yes, there are proofs, but then a heuristic optimization is applied to a provablysecure 14:03:35 scheme to get a nice speedup; or No, there are no proofs for AEZ itself, although 14:03:38 the authors employ provable-security to motivate and justify some design choices" 14:04:54 err… i suppose that was humour, yes :) 14:05:09 well, I thought it was funny 14:05:14 it made me puke in my mouth a little at the phrase "employ provable-security" 14:05:15 :) 14:05:59 really it should be "we proved that something else is secure! Alternatively, under a dubious counterfactual, this is provably secure!" 14:06:08 but it does sound quite funny now that i realise it's probably sarcasm 14:06:26 My guess is that they wrote just one of those sentences in an earlier draft 14:06:55 huh. I guessed wrong. 14:07:04 the v1 draft says: "Either: Yes, there are proofs, but then a heuristic optimization is applied to a provablysecure 14:07:07 scheme to get a nice speedup; or Yes, there are proofs, but under a nonstandard 14:07:08 assumption; or No, there are no proofs for AEZ itself, although the authors employ 14:07:11 provable-security to motivate and justify design choices." 14:08:45 ok, drifting. Thanks, all! 14:08:47 #endmeeting