22:06:24 <teor> #startmeeting Sponsor 55
22:06:24 <MeetBot> Meeting started Thu Jan 23 22:06:24 2020 UTC.  The chair is teor. Information about MeetBot at http://wiki.debian.org/MeetBot.
22:06:24 <MeetBot> Useful Commands: #action #agreed #help #info #idea #link #topic.
22:07:11 <teor> gaba said: Firs item is how we are going to divide the time
22:07:18 <nickm> so, for dividing time: I don't know how much of my time-budget is allocated here, but fwict teor, you have a better idea than I do how the project should break down.
22:07:31 <gaba> and the pad is https://pad.riseup.net/p/JSrYZgX3npU0v4XFp3Ih
22:07:47 <nickm> so it might make sense for me to follow your lead and ask you if you have pieces in mind for me to do?
22:07:50 <gaba> In relation with budget is for 1 full time dev
22:08:09 <teor> Yes, I do have a better idea of the breakdown right now
22:08:14 <teor> But I am hoping that we will all be working from the same proposals
22:08:23 <teor> And from the same set of tickets
22:08:39 <gaba> we could check with catalyst to see if they could also help
22:08:46 <nickm> hm. like, there is a queue for the sponsor, and we each take items from it?
22:09:24 <teor> I'm happy to be the main developer, but it's also ok for you to occasionally take items
22:09:39 <teor> And there will be some things you can do better than me
22:10:01 <nickm> ok
22:10:10 <gaba> teor: I think you should lead it and delegate items when needed
22:10:15 <teor> One of the risks of having only one developer is that they get stuck on things that are tricky for them, but easy for another developer
22:10:45 <teor> ahf also said that he was interested in helping
22:10:57 <gaba> In than sense communicate to the team and/or nick when you think you are stuck is good.
22:11:18 <nickm> I'm glad to preemptively take stuff that looks interesting to me, or wait for you to ask me to do something; whichever you prefer
22:11:45 <teor> I don't always realise when someone else would find things easier to do
22:11:46 <nickm> I do need to know what fraction of my time to plan to spend on it, but that's something we can adjust as we go along.
22:11:55 <teor> I'm happy to start by delegating
22:12:28 <nickm> if we're planning to do fortnightly meetings, we could do a quick overview of what's coming up at each one, and people could list what looks hard among the upcoming tasks
22:12:30 <teor> But I want at least one more developer reviewing what I do every day, so if there are things that are easy fixes, they can comment and help out
22:12:47 <nickm> right; I am happy to be one such developer but I should not be the only one
22:12:48 <teor> I don't always realise when things would be easier for other people
22:13:05 <gaba> ok
22:13:13 <teor> We could ask ahf to also have a look at progress, since that's part of his role already
22:13:18 <gaba> yes
22:13:27 <gaba> next is how we are going to communicate about this projct
22:13:57 <teor> I think it depends on the stage of the project
22:14:15 <gaba> it is a very short project (6.25 months)
22:14:19 <teor> Initially, I will send ~3 proposals to the tor-dev mailing list, and other developers will review them
22:14:26 <gaba> I would do a monthly meeting to check on status
22:14:34 <teor> Then I expect we will use the ticket tracker and IRC day to day
22:14:45 <gaba> yes
22:14:49 <gaba> and status during our weekly meeting
22:15:04 <nickm> let's plan for email discussion of the proposals as they come out, with a possible followup meeting one week later if we don't have consensus
22:15:30 <teor> I can give status at each weekly meeting, but I can only discuss it at 1/4 meetings
22:15:39 <nickm> How would you feel about a sync time for people on this project more frequently than once a month?
22:15:50 <nickm> I'd be in favor personally
22:15:51 <teor> I would appreciate a weekly sync time
22:15:56 <gaba> ok
22:16:01 <teor> Even if it's just a very short sync
22:16:05 <gaba> If you both are ok with that I'm fine with weekly sync
22:16:08 <gaba> yes
22:16:10 <nickm> who should be there for the.  you, me, and...?
22:16:15 <nickm> *for that
22:16:45 <gaba> we can set a weekly time and announce it in case other people want to attend
22:16:48 <teor> You and me, and depending on the week, gaba might want to be there
22:16:53 <teor> Others can attend as necessary
22:16:55 <nickm> ok
22:17:16 <gaba> does this time work for both of you?
22:17:19 <gaba> on Thursdays?
22:17:29 <nickm> this time is hard for me; it's my family's regular dinnertime
22:17:40 <teor> I prefer 2300 UTC
22:17:42 <gaba> later or earlier?
22:17:44 <gaba> ok
22:18:01 <gaba> Thursday 2300 UTC
22:18:33 <nickm> that's 6pm for me; it still makes dinner hard.  could we go to 0000 friday ?
22:18:44 <nickm> (7pm thursday for me)
22:18:54 <gaba> that works for me
22:19:12 <nickm> is that okay for you teor?
22:19:13 <teor> Oh sorry can we not do Thurday/Friday, it's my Friday, and I'm not always at work Fridays
22:19:29 <nickm> I'm fine with another day.
22:19:40 <teor> Can we do 0000 Wednesday instead? Right after the patch party?
22:19:54 <gaba> ok
22:20:18 <nickm> 0000 wednesday -- 7pm my tuesday?
22:20:40 <nickm> i could do that
22:20:47 <gaba> Sounds good to me.
22:21:02 <teor> Oh sorry, um no
22:21:28 <teor> 0000 Thursday which is 1000 my Thursday, right after the patch party
22:21:41 <teor> Sorry, I got confused about midnight
22:21:50 <gaba> what time and when in UTC ?
22:22:02 <teor> 0000 UTC Thirsday
22:22:57 <nickm> ==7pm eastern on wednesday, I think
22:23:05 <nickm> I can do that time too
22:23:27 <nickm> I might need to schedule once daylight savings time starts
22:23:28 <teor> Yes, that time
22:23:45 <gaba> ok
22:23:53 <gaba> Thursday 0000 UTC
22:23:58 <teor> And yes, that's fine, it's only a short sync, we can change the time as needed
22:24:04 <nickm> ok
22:24:09 <gaba> ok
22:24:15 <nickm> so what's next to coordinate? :)
22:24:26 * nickm is about to go slightly higher latency; running back and forth to kitchen
22:24:39 <gaba> next item is tracking of the project. This is not something you have to do but so you know that I have this documents I update for me to follow how this is going
22:24:51 <gaba> First, we said we are going to create one ticket per objective
22:26:16 <gaba> and we have some KPI that we need to write down at the beginning and end of the project.
22:26:42 <teor> ok
22:27:25 <gaba> should we move to the specific activities with the plan you wrote teor?
22:27:40 <teor> I'd like to discuss the total estimates first
22:27:53 <teor> * 26 days network team
22:27:54 <teor> * optional 11 days network team
22:27:54 <teor> * 5 days metrics team
22:28:11 <teor> How does that compare to the actual time available with the funding?
22:29:14 <gaba> that is much less that what we have fund
22:29:26 <gaba> we have 6.25 months of 1 full time dev
22:29:29 <gaba> from the network team
22:30:05 <gaba> in the timeline document there is one sheet that is the original budget
22:30:09 <teor> Ok, so that's roughly 130 days?
22:30:11 <gaba> it says how much time we have for each obj
22:30:12 <gaba> ttps://nc.torproject.net/s/XrsgxQDdkBq3SeF
22:30:23 <gaba> Not really. I'm calculating 3 days per week
22:30:43 <gaba> more around 78 days
22:32:13 <teor> Ok, I am very happy with a 3x under-run in my initial estimates
22:32:39 <nickm> i bet we will find plenty of things to do
22:33:00 <teor> I already think I may have under-estimated by at least 1.5x-2x, because the proposal has taken 2-3 days to write already, and I estimated 2
22:33:22 <gaba> yes, estimation is hard
22:33:25 <teor> There is plenty of optional work listed on the pad
22:33:48 <teor> As I write the proposals, I'm going to split the optional work into "should" and "can" items
22:33:57 <gaba> ok
22:34:31 <teor> the should items have a significant impact on security, reliability, or efficiency
22:34:51 <gaba> in the timeline document I tried to translate/understand your pad into one of the sheets. Everything is estimated there. We could double the estimations if you didn't do it already.
22:35:01 <gaba> All of them needs tickets.
22:35:28 <teor> Not yet, please
22:35:30 <gaba> Maybe we could go through all the objectives and see if Nick is ok with the plan for each of them?
22:35:51 <teor> I want to write the proposals, then make tickets
22:36:22 <teor> I am trying to edit the plan as I write the proposals, but I would like to work off the proposals to make tickets
22:36:57 <gaba> ok
22:37:04 <gaba> what do you do in this sync/meeting now?
22:37:09 <teor> I am discovering missing work and unnecessary work as I write down the details in the proposals
22:37:14 <gaba> what do you want to do* in this sync/meeting now?
22:37:18 <gaba> sorry, I'm missing words
22:37:31 <nickm> it seems to me that we might be better off making high-level tickets for the high-level objectives---
22:37:41 <nickm> and then making sub-tickets for individual items as we approach them?
22:37:48 <gaba> yes
22:37:59 <teor> Yes, the objective tickets are for gaba, and she can make them whenever she likes :-)
22:38:01 <gaba> that is what I meant by 'one ticket per objective in Trac'
22:38:16 <gaba> teor: good! I didn't know if you wanted to do them
22:38:29 <teor> And yes, I think we should make developer tickets as we do the work
22:38:30 <nickm> and the proposal-based tickets should be child tickets of the objective tickets?
22:38:35 <teor> Yes
22:38:36 <nickm> sounds good to me if so
22:38:38 <gaba> yes
22:39:44 <teor> We'll need to split some of the tickets from the proposals into objectives, because there are 3 proposals, and 5 objectives
22:40:53 <teor> So I'd like to give an overview of the proposals I'm writing, and how they map to the objectives
22:41:27 <teor> Proposal 311 covers Relay IPv6 Extends and Relay IPv6 ORPort Reachability Checks
22:41:50 <teor> It maps to objective 1.1: Propose and implement IPv6 ORPort reachability checks on relays
22:41:54 <nickm> are any protocol changes needed here, or just behavior changes?
22:42:18 <teor> and the chutney testing maps to O1.3 Integration test Tor relays over IPv6 using chutney
22:42:50 <teor> nickm: one moment, just reading the proposal draft to check
22:44:36 <teor> Relays will use the existing EXTEND2 cells to send IPv6 ORPort information, and other relays will need start extending over IPv6 in response to those EXTEND2 cells
22:45:26 <teor> So we don't need to change cell formats, but we do need to change remote relay IPv6 connection and circuit extend behaviour, as well as the reachability tests on the bootstrapping relay
22:45:47 <nickm> ack
22:46:11 <nickm> makes sense
22:46:20 <teor> Because the bootstrapping relay needs to select relays with the new behaviour, relays will need to advertise a new Relay subprotocol version
22:46:59 <nickm> hm, ok.
22:47:46 <teor> I can't think of any other way to do it efficiently as we deploy, but let's discuss when I send the proposal draft?
22:48:58 <teor> We only have 10 minutes left, so I'm going to move on to the next proposal
22:49:32 <nickm> +1
22:49:47 <teor> Proposal 312 covers Automatic Relay IPv6 Address Detection
22:50:06 <teor> It maps to O1.2 Make relays figure out their own IPv6 address
22:50:16 <teor> and the chutney testing maps to O1.3 Integration test Tor relays over IPv6 using chutney
22:50:43 <teor> There are no protocol changes required for this proposal
22:51:08 <teor> But relays will start fetching some directory documents over IPv6 from IPv6 authorities
22:51:43 <teor> (Once they've tried all their local options for finding their IPv6 address)
22:52:30 <nickm> so that leaevs O1.4 and O1.5?
22:52:49 <nickm> *leaves
22:52:54 <teor> O1.4 belongs to metrics, it's an occasional analysis
22:53:03 <nickm> ack
22:53:10 <teor> O1.4 Measure the number of Tor relays that support IPv6 reachability checks
22:53:20 <teor> O1.5 Measure the number of connections, and consumed bandwidth, using IPv4 and IPv6
22:53:43 <teor> maps to a proposal about IPv4 and IPv6 connection and bandwidth statistics
22:54:09 <gaba> We need tickets for all this stuff that needs to happen in each OX.X activity.
22:54:28 <gaba> After that we can include this tickets in the metrics team roadmap.
22:54:39 <teor> I'd like metrics to create the tickets for O1.4 Measure the number of Tor relays that support IPv6 reachability checks
22:55:10 <teor> And for their parts of O1.5 Measure the number of connections, and consumed bandwidth, using IPv4 and IPv6
22:56:27 * nickm summing up since I have to run in 2-4 minutes
22:56:28 <gaba> karsten may need some help on this. Let's check with him as I had the impression you were going to create the tickets for this activities.
22:56:35 * gaba needs to run soon too
22:56:57 <teor> We can also put some IPv6 statistics in Tor's heartbeat logs as part of O1.5, but that's optional
22:57:08 <nickm> it sounds like our next steps here are the proposals.  If they're done next week, excellent.  But if you think it would help, you should also feel free to show me the drafts with any open questions that you have, and I can try to help with them
22:57:27 <nickm> (sorry for interrupting, my brain is context-switching hard)
22:58:01 <teor> gaba: If you want me to make metrics tickets, please let me know which objectives, whether I should include optional work, and what date you need the tickets to be done by
22:58:43 <gaba> teor: I would like you to make the tickets that are missing from the activities you added in the pad. Metrics team can review them later.
22:58:58 <teor> nickm: I'd like to finish up the first draft of the first proposal today, including some open questions
22:59:03 <gaba> I will do the tickets for the objectives and we attach the ones you already know we need to have for each obj.
22:59:13 <teor> gaba: whether I should include optional work, and what date you need the tickets to be done by
22:59:36 <nickm> okay; I'll look at it when I start in the morning.  I've allocatd a few hours to my friday for this
22:59:38 <gaba> By next week before we start the project would be the best.
23:00:12 <teor> Sure, please let me know when you've made tickets for each objective
23:00:16 <gaba> ok
23:00:31 <teor> to confirm: I will not make tickets for the optional work yet
23:00:42 <nickm> sounds reasonable to me
23:00:49 <gaba> sounds good
23:01:02 * nickm needs to go offline; my apologies.  Thanks so much to teor and gaba for your work on this!
23:01:09 <nickm> I'll check backlog before I go to bed
23:01:30 <teor> Also to confirm: I'm only making metrics tickets right now. Network team tickets need to wait for the proposals.
23:01:49 <gaba> teor: we can not make tickets for the activities you are proposing in the pad?
23:01:56 <gaba> nickm o/
23:02:27 * gaba is a little confused with all the letters/lines and activities then
23:02:47 <teor> gaba: as I said earlier: I am trying to edit the plan as I write the proposals, but I would like to work off the proposals to make tickets
23:02:54 <gaba> ahh, ok
23:03:22 <teor> I don't want to have to keep a pad and the proposals and the tickets in sync, the work duplication doesn't make sense
23:03:31 <gaba> nop, this pad is temporal
23:03:36 <gaba> the idea is to move everything into trac
23:04:40 <teor> When the proposals are done, I'll check that every item in the pad is in a proposal, or in a list of extra tickets, or I decided not to do it
23:05:01 <gaba> ok. I need to go too
23:05:21 <gaba> let me know when you are done with the proposals next week
23:05:50 <teor> I'll send them to tor-dev as drafts, we will make tickets after nickm and others have reviewed the proposals
23:06:00 <teor> I can't guarantee they will all be done next week
23:06:23 <gaba> bye! o/
23:06:34 <teor> bye :-)
23:06:35 <teor> #endmeeting