Wednesday, 2018-03-21

*** annabelleB has quit IRC00:59
*** diablo_rojo has quit IRC01:04
fungitc-members: looks like time for another office hour01:06
fungithough i'm already a bit sleepy01:07
*** harlowja has quit IRC01:07
*** gcb has joined #openstack-tc01:08
* persia wonders if 9 or 10am is usually a good time for governance discussions for folk in UTC+8 and UTC+901:49
fungino clue, but tends to be a ghost town for us at any rate01:58
fungiwe also have 09:00z on tuesdays for apac night-owls i suppose02:01
fungi(5 or 6pm in your mentioned timezones)02:01
fungianyway, domestic responsibilities call me away now that this particular office hour has concluded02:02
*** mriedem has quit IRC02:06
*** rosmaita has quit IRC02:53
*** harlowja has joined #openstack-tc04:39
*** harlowja has quit IRC04:56
*** kumarmn has joined #openstack-tc06:36
*** kumarmn has quit IRC06:39
*** kumarmn has joined #openstack-tc06:50
*** kumarmn has quit IRC07:05
*** diablo_rojo has joined #openstack-tc07:34
*** csatari has joined #openstack-tc07:53
*** amrith_ has joined #openstack-tc08:37
*** amrith has quit IRC08:37
*** amrith_ is now known as amrith08:37
*** jpich has joined #openstack-tc09:02
*** amrith has quit IRC09:05
*** kumarmn has joined #openstack-tc09:06
*** amrith has joined #openstack-tc09:06
*** amrith is now known as Guest5087709:06
*** Guest50877 has quit IRC09:07
*** amrith_ has joined #openstack-tc09:08
*** amrith_ is now known as amrith09:08
*** kumarmn has quit IRC09:10
*** amrith has quit IRC09:12
*** amrith has joined #openstack-tc09:13
*** amrith is now known as Guest2199009:13
*** Guest21990 has quit IRC09:14
*** amrith_ has joined #openstack-tc09:15
*** amrith_ is now known as amrith09:15
*** amrith has quit IRC09:22
*** amrith has joined #openstack-tc09:23
*** amrith is now known as Guest5681709:24
*** Guest56817 has quit IRC09:28
*** amrith has joined #openstack-tc09:49
*** amrith_ has joined #openstack-tc09:49
*** amrith has quit IRC09:49
*** amrith_ is now known as Guest6353509:49
*** Guest63535 is now known as amrith09:50
*** dtantsur|afk is now known as dtantsur10:49
*** cdent has joined #openstack-tc11:18
cdentmornin' dims11:57
*** rosmaita has joined #openstack-tc12:08
*** gcb has quit IRC12:42
*** rosmaita has quit IRC12:53
dhellmanngiven the recent conversation in the first-contact sig thread, I wonder if APAC users are going to be on IRC at all or if they'd be better served by having office hours on weechat or whatever other tool is more popular there13:12
cmurphyI think we should still be encouraging everyone to converge on the tools we already use, the big problem I think isn't the tools themselves but the languages we speak when using these tools13:17
cdentdhellmann: many of the people from china that I communicate with on IRC are only able to do so outside of their office, so I would agree: trying to enforce use of IRC that is unlikely to result in success13:17
cmurphythat is also a problem13:17
cmurphymaybe we should be trying to talk to the contributing companies to relax their firewall rules so that their contributors can get things done13:19
cdentbut also what cmurphy said: those that can use irc are likely struggling on the language front13:19
cdentI find irc to parse and I've been speaking english for way too long13:19
cdentapparently I also find it hard to write, because I left a word out: hard to parse13:19
*** mriedem has joined #openstack-tc13:20
mugsieand getting companies to relax these rules can be problematic ... I know I have no outbound IRC connectivity in the office, and never will, due to our data security and integrity rules.13:21
dhellmannI guess weechat is a client and wechat is a service? are they related? Put me on the "not hip enough to know this stuff" list13:22
cmurphythey are not related13:22
jrollweechat is an irc client :)13:22
cmurphyit's very confusing13:22
mugsieof course, out bound SSH / SSL connections are not blocked, so I can connect to my bouncer just fine, but ymmv13:22
dhellmannI guess I should turn in my Interneter's License13:22
jrollwechat, AIUI, is the chinese equivalent of whatsapp and similar13:22
cmurphyright, wechat is not at all common outside of China so it's reasonable not to have heard of it13:23
dhellmannI'm also wondering if anyone other than us gets any value out of the office hours. I mean, I do, for the ones where we actually have a conversation. But it seems like it's usually the same folks and we don't have many people coming with discussion topics from the broader community.13:23
cdentdhellmann: yeah, I've been thinking about that too.13:24
cmurphyi don't think that's true13:24
dhellmannmugsie : I've never understood those sorts of policies, but I do agree they're likely a barrier. Would they be less so if we were on a different platform, though?13:24
mugsieI think it has helped spread the frantic hour of interleaving conversations across the week13:24
cmurphywe had tbarron come here to bring up an important topic the other day13:25
cdentAnd I don't know. I think the conversations are useful, but they have the same costs as the original tc meeting13:25
mugsiedhellmann: nope. its a ban on all chat services13:25
cdentmugsie: yes13:25
jrollI often read office hours scrollback rather than participating. it's useful to me to be able to read that.13:25
jrolland I know rosmaita has brought a topic here at least once13:25
dhellmannI'm prepared to say that building a team requires paying some cost in terms of establishing good communication.13:26
dhellmannyeah, I'm not saying *no one* brings topics. It just seems less common than I think we hoped it would be.13:26
dhellmannand maybe people are choosing to bring topics up on the ML or we're identifying them ourselves instead of waiting13:27
dhellmannor maybe I had a different expectation for what we would be doing than others did13:27
jrolldid people often bring topics to the TC meeting, outside of governance reviews put on the agenda?13:27
persiaSo, "APAC" != "China".  IRC in China is tricky, but some of the tools used in China are not used elsewhere in Asia (and almost entirely unused in the antipodes)13:27
cdentdhellmann: can you expand on your expectation?13:27
cmurphypersia: excellent point13:28
dhellmannI thought the point of having them at different times was to reach other groups, though, and I wonder if they're (a) not interested (b) using a different platform or (c) both13:28
persiacmurphy: The "firewall rules" that apply in China aren't usually imposed by companies: most are imposed by the government.13:28
cdentlittle from column a, little from column d13:28
dhellmannjroll : that's a good question. Earlier on, I think so. More recently, I think less.13:28
dhellmannthat is, more recently up to the time we stopped having them13:29
jrolldhellmann: that's what I recall as well13:29
cmurphypersia: I think some of them are imposed by the companies, I've heard of contributors not being able to use IRC at the office but able to at home13:29
mugsieI would say a few people might have ethical issues using wechat13:29
dhellmannpersia : My point was that I have not seen much participation for the office hours time slot established to be convenient to APAC time zones regardless of country, and I wonder why that might be.13:30
persiacmurphy: An interesting question to ask in such a situation is whether they are doing so through some of the known holes in the "Great Firewall": some things are permitted (with careful monitoring) by individuals, but not permitted by organisations subject to audit (especially those with partial state ownership).13:30
mugsieand the releative quietness of the "APAC friendly" office hours where it can just be fungi talking into the void might not encourage people to bring things up, as there is only a small amount of TC members around?13:31
persiadhellmann: I agree with your point.  I think it is an interesting question.  I think saying "they are on weechat" is focused on a narrow set of folk in the +5:30 to +14 timezones.13:31
cdentdhellmann: just to try to keep the question moving along: in your head, what sorts of thing would you expect a non-tc person to show up to office hours for?13:31
dhellmannmugsie : sure. I think we started with the Field of Dreams assumption but given the low turnout from the TC itself I don't know if we ever did actually "build it"13:31
dhellmanncdent : that's a good question, too.13:32
cmurphypersia: the other thing I've heard of is restrictions on sending external email from within the company, which wouldn't necessarily be due to a firewall rule but a policy enforced by the corporate mail server plus unwillingness or inability to use personal email in the office13:32
dhellmannthe interop testing qa/designate/heat thing might have come up that way, if that had been handled better13:33
persiacmurphy: Yes.  The restrictions on email are a corporate thing.  I have never heard of state-mandated restrictions about that (although my information may not be complete).13:33
cdentcmurphy: yup, i've heard that one too13:33
mugsiedhellmann: it did initially :)13:34
dhellmanncmurphy, persia : I think we're going to have to differentiate between people not being allowed to do something and us not making it possible for them to. Companies that refuse to allow anyone to send email to our mailing list are not going to have good interactions with the community and we can't necessarily fix that. If we fail to support communication channels that people *are* allowed to use, we can examine and13:35
dhellmann potentially change our own behavior.13:35
dhellmannmugsie : succes? :-)13:36
mugsiedhellmann: well, we saw what happened  ... so no.13:36
dhellmannwell, no13:36
dhellmannthere is plenty of blame to go around in that, including to the TC13:37
dhellmannand before I start ranting about that again, I'm going to go make breakfast13:37
persiadhellmann: Yes.  I think a baseline assumption for any contributing organisation is that they permit their involved staff to use a consistent email address independent of location (to deal with odd hours imposed by our community which apply regardless of timezone: just ask the UTC-5 to UTC-8 people who start early), and permit access to our git repositories, issue trackers, etherpads, wiki, etc.13:37
persiaThat might need to be documented in a "How your organisation should contribute to OpenStack" document13:38
persiaWhere there are state-imposed restrictions, we can do less.13:38
dhellmannsome of that is covered in13:38
persia(although we've done some, like relaxing the requirement for https to use our web resources, as at least one state doesn't permit international https, due to the difficulty in scanning it for inappropriate content)13:38
persiaThe topics are covered, but the document appears written for an audience of individuals, so folk looking for corporate involvement may find things different.13:39
mugsiepersia: I think the feedback needs to come from these companies about what is state sponsored issues, and what is internal company policies13:39
* mugsie mutters something about how we have many foundation members in countries that have these issues, and if they were serious about doing the community bit they would have been shouting about it by now13:40
cmurphymugsie: that puzzles me too13:41
persiamugsie: Perhaps.  I don't think they will do that unsolicited.  I also know of companies in the US and UK who have policies about personal chat/email that would block cooperation, some of which might match state policies in some states (but not those).13:41
persiaAlso, it is easier to shout "I'm being discriminated against: do special things for me" than "Go fix the problems, I don't want special treatment".13:41
mugsiecmurphy: I wrote about this last week, but look at the contributions from one of our latest gold member -
*** dklyle has joined #openstack-tc13:44
dmsimard(I'm late but re: no one brings topics during office hours) I tend to just ask things here whenever something comes up and my expectation is that I'll get replied to at some point even if it's in a few hours. My schedule/hours are pretty hectic so sometimes it's hard to say, like, I'll write this down for the next office hour and hopefully I 1) won't forget about it 2) I'll be able to attend13:44
cmurphymugsie: wow :(13:44
*** david-lyle has quit IRC13:44
mugsieyeah :(13:45
dmsimardmugsie: I thought sponsoring companies were mandated to commit (human) resources to the project ?13:46
mugsiedmsimard: they are13:46
mugsie has a brief overview of my thoughts13:46
cmurphy"mandated" but no one enforces it13:46
* cdent mutters (late) along with mugsie13:47
dhellmannI thought the mandate only applied at platinum level13:48
mugsieone of the good things I have seen from the office hours is this channel - people like dmsimard amrith rosmita have all dropped questions in here outside of office hours which has caused discussions, and  allowed discussions like this one to spontainiously spark to life13:48
persiaI can't find any evidence of that from
dhellmannthere used to be a policy that platinum sponsorship had to include $$ and "2 full time equivalent" contributors13:50
dhellmannI don't know where that was written down, though13:50
persiaI remember that statement.  I can't find it today.13:50
dmsimardpersia: "For reference, the Platinum Members each contribute $500k USD per year to the foundation and must also have the equivalent of 2 full time employees contributing to OpenStack."13:50
cdentmugsie: yeah, I definitely like the existence of the channel. It's _so_ much easier to be able (and want) to participate because it is more focused13:50
persiadmsimard: Thank you.13:50
dmsimardNo mention of human resources for gold as far as I can see13:51
dhellmannyes, good point on the fact that folks do ask questions at any point here in channel13:51
mugsieit was discussed in the session in sydney when tencent were added, but I do not know the basis of it13:51
mugsieand we (non directors) were kicked out fairly soon after for the executive session13:52
persia"contributing" also is not well defined, they are not required to be ATC, for example.13:52
cdentwhatever the law, the spirit isn't really being followed, especially given the discussion about contribution that happens whenever a gold member applies13:53
zaneb2 FTEs is such a low bar (that some platinum members still fail to meet) that there hardly seems any point even having a bar for gold members13:53
dhellmannpersia : good point. i wonder if they are contributing in other ways13:53
cdentzaneb: didn't you know, there are really only 16 people working to create the multi-billion dollar market that is openstack13:53
mugsiewell, <cynicism> they are paying the foundation, and letting a real nice logo be on the website </cynicism>13:54
dhellmannand I guess the other side of the equation is, are they expecting more from the community than they are willing to give?13:54
zanebcdent: that sounds about right ;)13:54
dhellmannI'm not aware of any huge demands for features from tencent, in particular.13:54
persiamugsie: They may also be doing things like assigning someone to be their "OpenStack Programme Manager", and someone else to be their "OpenStack Ambassador", both of which people may have frantic more-than-full-time jobs, even if their activity is not that visible outside the organisation.13:55
persiaBut I think looking to the rules for contributors is not to anyone's advantage.  OpenStack would do poorly with 16 ATCs.13:55
dhellmannappointment time, bbiab13:57
*** zhipeng has joined #openstack-tc13:57
mugsiepersia: Oh, I know ... but it is not unreasonable for an expectiation of someone joining a foundation to be an active member in the community, and for the board to consider how will this new member help further the goal of building first class software13:58
mugsieand neither an internal programme manager or ambassador do that. they help the company use / integrate / build on top of the software , but those jobs would have been there if they joined or not13:59
mugsiehonestly, short of being able to say you are a member, I am not sure why you join if you were not helping to shape the software itself14:00
fungicatching up... the technical resources companies need to provide to people they assign to contribute to the project may be something diablo_rojo wants to add to the agenda for the next meeting of the first contact sig (once she wakes up from running the last one 6 hours ago)14:00
mugsie(helping to shape the software covers a lot more than dev)14:00
mugsiefungi: oh, good call.14:00
fungihad a nice geographic spread of attendees too, granted they were only 414:00
zanebpersia: that's true, but there's something to be said for the fact that getting your first contributor set up is the hardest. as discussed above, you have to change your policies, modify your firewalls, &c. If we had a mechanism to encourage people to get over that hurdle then maybe getting additional contributors would be easier14:01
fungii have trouble convincing myself to attend a 0800z meeting regularly, but have been making a point of reading the meeting logs after and it seems like they're picking up steam14:01
dmsimardzaneb: at the same time, if you're going to commit upwards of 500k$ to the foundation, should you not have some amount of competence in OpenStack before doing that ? Implying knowledgeable contributors would hopefully already be a thing14:02
dmsimardThere's definitely challenges in getting started prior to sponsoring if you've never touched OpenStack before14:03
zanebdmsimard: you'd think so but it's their money I guess :D14:04
fungiin the tencent case, it may also be possible that we have bad data. stackalytics isn't in the best of shape, it expects people to contribute configuration to it so that it can map their e-mail addresses to organizations, and that configuration has often gone months without getting reviewed in the past. i wouldn't put a lot of stock in those analyses necessarily14:04
fungiit also doesn't pick up configuration changes without a restart and full reindex14:05
dmsimardzaneb: I suppose some companies have absurd amount of disposable pocket change :)14:05
fungi(we learned this when trying to start running a copy of it within the project infrastructure)14:05
cmurphyfungi: how do we encourage companies to provide those technical resources for their contributors? irc and email are the biggest barrier and probably the hardest corporate policies to change14:06
*** kumarmn has joined #openstack-tc14:06
dmsimardfungi: I remember seeing some kind of contribution statistics for k8s projects.. it was in grafana perhaps ? I wonder how they pull their data.14:07
fungicmurphy: i agree. for many it may be as simple as giving your employee a cloud account and making an exception that they can ssh or vpn to it during work hours14:07
fungii mean, we discovered that there were also plenty of companies who block the non-https api interface to gerrit because it runs on an odd high-numbered port14:07
fungiand contributors in most cases found workarounds to that in years past14:08
cdentWe're probably better of not speculating. Instead we can probably ask some people. Who are those people?14:08
dmsimardfungi: Ah, there it is:
*** hongbin has joined #openstack-tc14:08
fungi"we don't know because they can't e-mail us or tell us in irc" ;)14:08
fungidmsimard: yeah, bitergia also used to make the site for us, but we never had luck explaining our contribution model to them in ways they could understand enough to accurately model14:10
cdentwe have board members from the companies in question, yeah?14:10
dmsimardmugsie: ah that's actually the one I was looking for -- the is one I hadn't seen before14:10
mugsiedmsimard: I personally dislike it, but you can dig in to the stats quite well if you know graphana14:10
*** kumarmn has quit IRC14:11
dmsimardmugsie: oh, I was mostly pointing those out because it seems like they attribute commits to organizations -- on the topic of stackalytics not being fully accurate due to affiliation remapping with emails14:11
fungibitergia is a company whose business is analyzing development communities (both open and proprietary) and selling them tailored analysis and reporting tools14:11
dmsimardI was curious how they ended up attributing commits to companies -- if it was just about email addresses or if it was something else.14:11
mugsieI think it comes from the linux foundation CLA ?14:12
fungiyeah, we have a similar lookup api for foundation member profiles, but it postdates when stackalytics development was all but abandoned by mirantis or it would have possibly made use of that14:13
mugsiecdent: yes, from looking at we have directors from companies in that region, but not from the specific example I used above14:13
cdentmugsie: yeah, what I'm suggesting is that people in that region probably know better than us what's happening in that regiion14:13
*** annabelleB has joined #openstack-tc14:14
mugsieyeah - potentially a good item for the joint BoD / TC session in YVR (if that is happening again)14:14
mugsiewhich it probably should seen as the Kata and Zuul people will have the first opportunity to join a joint leadership session14:15
cdentwe're going to need a bigger boat14:16
fungifwiw, the "zuul people" were already showing up to those anyway and regularly dial into the remote board meetings too ;)14:16
mugsiefungi: yeah - I know, but this will be the first time they will be "Zuul People" :)14:17
* fungi is pretty sure they were already zuul people14:17
smcginnisTBF, the idea of foundation member companies being able to provide a "narrative" of their contributions to OpenStack was brought up partly to get past the idea that "contributing" had to be something that shows up in stackalytics.14:18
fungi(at least once they started writing zuul in 2012)14:18
smcginnisAnd way late now, but there are some companies in China that do have access to more past the Great Firewall than normal citizens can get from home. But that means they need to be on their work network to get to those things. And it is probably pretty limited.14:19
smcginnisEven in the US, at $EMPLOYER-1 I had to submit a yearly "security exception request" in order to get IRC ports open, which involved giving our customer demand for our involvement in OpenStack as a justification to allow it.14:20
persiafungi: "zuul people" vs. "Zuul People" is the distinction there.  In the first case, it is a subclass of "OpenStack People", in the second class, not so much.14:20
mugsiefungi: what persia said :) - I just meant that as we grow stratigic focus areas we need to figure out things like joint leadership sessions and how we deal with them getting bigger14:22
fungisure, i agree we're gonna need a bigger boat, just pointing out that zuul people were already in the boat we have14:26
fungiit's more of a governance refactoring in that particular case, not a bunch of new people coming to us looking for more seats in the boat14:27
fungikata on the other hand, yes, new people, need to find somewhere they can sit comfortably14:28
persiaMy thought is that all of "Datacenter Cloud Infrastructure", "Containers Infrastructure", "Edge Computing Infrastructure", and "CI/CD Infrastructure" should have a time allocation similar to the previous BD/TC meetings.  This can't actually happen in a single day, which means that folk get squeezed or the format changes.14:32
* TheJulia reads14:32
persiaThis isn't just about how many people are in the room, but more about how much time is allocated for the topics of those people.  There was usually a quick update on zuulish things as part of the BD/TC meetings, but if that is a strategic focus, I would hope both the board and the "Zuul People" take more time to consider things with more granularity.14:33
fungior we come up with a new discussion format at those meetings which isn't divided up by focus area14:33
persiaThat also works, as long as everyone has a chance to discuss their issues in enough depth to depart the meeting with confidence.14:33
funginoone ever has a chance to discuss their issues in that much depth, from what i've seen14:34
persiaI have favorable memories from austin, barcelona, boston, but perhaps that is true.  I remember being confused in both denver and dublin and missed sydney.14:36
* ttx reads14:48
ttxCouple of notes: wechat in China is not just "their local variant of Whatsapp", it basically replaced almost all forms of communications there14:49
* smcginnis can attest to that14:49
TheJuliafungi: A new discussion format sounds ideal, and as persia pointed out that there is value in providing time to discuss items in depth. The thing I'm wary of is going down the rabbit hole on issues that do not have overlapping reason for being on an agenda for said format.14:50
TheJuliattx: including two people sitting in a room, or just general electronic public communication?14:52
ttxit also replaced phone calls, SMS and payment14:52
ttxRe: the contribution requirements for Platinum membership, they don't appear in the bylaws or the platinum member policy14:53
persiaTheJulia: It replaces people sitting in a room they way that other electronic forms have replaced voice in other countries.  Consider the case of folk sitting next to each other texting each other, etc.14:53
ttxbasically IRC or Slack or whatever will always be seen as openstack-specific, so as long as it's not blocked I think they don't really care14:56
* TheJulia wonders if there is really even sufficient social context into that culture, such that the real thing to do is to learn more14:57
persiattx: I think that's a little extreme: there are other communities that use $whatever, and some organisations may be involved in them.14:57
TheJuliapersia: but is the their org facilitating? or is it happening naturally?14:58
persiaTheJulia: The real thing is to learn more.  The middle kingdom is special in many ways that are opaque to even folk who live in areas previously belonging to that hegemony.14:58
ttxsure, my point is more... they don't have that much of a strong preference in anything else than wechat :)14:58
persiattx: Totally agreed on that point, as long as we say "China" and not "APAC" :)14:58
ttxYes i highly recommend spending a few days there at least once14:58
ttxif you get a chance14:59
persiaNote that spending time in Taiwan, Hong Kong, Chinese portions of Malaysia, etc. does not count.14:59
ttxI would recommend going all in and visit Beijing14:59
persiaGuangzhou or Shanghai might work as well.14:59
pabelangerheads up, I sent out reminder on poll for making release S. it will be closing later today.15:00
ttxpabelanger: can we pay you to get early results ?15:03
pabelangerI mean, you could. But I don't think I'd release them :)15:05
pabelangerTBH, I haven't looked myself15:05
zhipengShenzhen as well :)15:19
*** zhipeng has quit IRC15:26
dtroyerFWIW, I recall there being a 1FTE requirement for Gold members in addition to the 2FTE requirememnt for platinum members in the early days, apparently none of that written down, nor ever enforces to my knowledge.  Complaint about members not fulfilling that requirement are almost as old as the Foundation.15:27
*** kumarmn has joined #openstack-tc15:32
TheJuliaAnd often people forget to put themselves in other people's shoes since we don't walk the same path, we just so happen to walk similar paths15:33
dmsimardcdent: unrelated but I was reading your TC summary and it would've probably been fair to include the announcement that Jim eventually sent about Zuul ( )15:48
cdentdmsimard: a) yes, you're probably right, b) I make zero promises about being fair15:49
cdentdmsimard: please feel free to followup with a response to the list and/or a comment on the blog15:49
dmsimardMostly making sure you didn't miss that particular email in a flood of openstack-dev emails :)15:50
cdentYeah, I saw it.15:50
cdentAs I tried to make clear in my comments on the topic in discussion here: I actually think the "sin" about zuul is _very_ minor. What's a problem is that at least some people feel it is an example of a fairly common failing that we (whoever that is) make.15:51
fungii expect the fact that nobody actually followed up to that announcement expressing surprise indicates that 1. everyone was already aware because we've been announcing this in bits and pieces all along, 2. nobody who didn't know cared one way or the other, or 3. the people who are likely to be surprised by this also don't read the openstack-dev mailing list15:54
cdentfungi: you realize that's entirely not the point?15:55
fungithen i'm still not sure what the point is. we had already done our best to reach out directly to people who had a vested interest in the situation early in the process15:56
fungii mean, there have been keynotes on stage, discussions in board and joint leadership meetings, mention in infra weekly meetings, tc office hours, on the infra mailing list, in person at the ptg...15:57
fungii don't object to mentioning it yet one more time but the governance change proposal was really the next step in communication15:57
*** rosmaita has joined #openstack-tc15:58
cdent[t We5]15:58
purplerbot<cdent> As I tried to make clear in my comments on the topic in discussion here: I actually think the "sin" about zuul is _very_ minor. What's a problem is that at least some people feel it is an example of a fairly common failing that we (whoever that is) make. [2018-03-21 15:51:06.853165] [n We5]15:58
cdentthat "common" thing came out in the discussion at last Thursday's office hours15:59
funginot sure which common failing you mean there. we all have plenty of those i guess15:59
fungithere were many concerns raised in that discussion but the primary one seemed to be that we needed another mailing list thread, i thought15:59
cdentthe fact that we went on about it for such great length suggests people experience issues with transparency and communication and that _despite_ all the wonderful things that infra-folk did to prepare, it still caused some consternation16:00
cdentIt was that we need to get better about not needing to be reminded that we need another mailing list thread16:01
fungiahh, got it. yes i do agree that we already do a fine job of (perhaps even over)communicating, but for whatever reason people who _aren't_ at all surprised by things continue to raise concerns that they think there might be people somewhere who care abotu these things and haven't heard16:01
* cdent glances askance at fungi16:02
fungiand it's time for us to accept that we've come up with lots of ways to communicate effectively to people who are receptive to communication, and move on to other more pressing concerns16:02
cdentwe are inconsistent and can do better16:02
cdentwhich is not the same as "we do a bad job"16:03
cdentwe don't do a bad job16:03
fungiand our efforts may be better spent on other things we do a worse job on16:03
dmsimardoff-topic: what's that purplerbot thing ?16:04
smcginniscdent: Sorry, had it handy.16:06
cdentfungi: you may be right, but I'd argue that we'll struggle to reach agreement on what it is that we're doign a worse job on until we've got some of these underlying issues resolved16:07
cdentmore right is probably some middle of the road16:08
persiaNote that communications are slippery.  I can say that there are lists to which I neither subscribe nor read archives intentionally, as not knowing about certain things allows me to make statements in good faith without regard to reality as perceived by other folk.  I don't think I'm the only person who does that sort of thing (and actually, suspect I'm more likely to have encountered communications than many folk who deal with marketing, business16:08
persia relations, and other areas where ignorance can be a source of power).16:08
persiaThere may also be folk who are made aware of communications but choose not to share that information (or are otherwise not careful about provenance of information of consistency of narratives)16:09
persias/information of consistency/information or consistency/16:10
dmsimardcdent: Maybe it boils down to interest ? People interested in Zuul most likely knew about it. If the issue is that people not interested in Zuul did not know about it, I'm not sure what communication channels would be best and if it would even reach that audience (because either they don't care or they're not interested). I mention Zuul but as you said, we can replace Zuul with $thing.16:10
dmsimardLike if $project does a backwards incompatible change or something and announces it widely on openstack-dev -- I might not be subscribed to the threads for $project because it's not relevant to me and I'm cutting down on the email volume16:12
fungiin this case the governance change was an incremental step in a much larger effort. the larger effort has been communicated plenty of places and the governance change itself was not particularly more or less significant than other times an existing team officially drops responsibility for one or more of its deliverables16:12
cdentsigh. how many times do I need to say "it's not about zuul". People expressed surprise about a thing that happened. A role that I've taken on as a TC member is trying to improve transparency and participation. If I hear people expressing that they feel like they missed out knowing something, then I want to amplify that concern.16:12
fungii don't recall anyone expressing surprise in this case16:13
dmsimardcdent: as I mentioned my statements can s/Zuul/$thing//16:13
fungii recall people raising concerns that some people might be surprised. i don't think anyone actually _was_ surprised, but it's hard to know. my guess is that the people who are likely to be surprised by it still haven't found out and a post to openstack-dev doesn't really change that16:14
cdentdmsimard: same response: any individual instance is not relevant (to me), it is the desire to amplify concerns about info sharing16:14
fungiand anyway, i have a hard time seeing how it's not about the particular governance change which dtarted the discussion... the circumstances of each change are unique in _some_ way and require people making judgement calls on where, when and to what extent they need to be announced/discussed both before and after proposing16:16
dmsimardcdent: ok, let me try another example. I read the TC summaries that you and ttx write up because I'm interested in them and what the TC is doing but I don't genuinely expect everyone (dev/user/operator/integrator) to read those or be interested in them.16:16
cdentdmsimard: yes, agreed16:16
cdentdmsimard: but what those things do is provide a regular, consistent, habitual form of info sharing16:16
dmsimardSo while I agree that more communication is generally better, it's not going to help if people aren't interested or don't care16:17
cdentthe discussion last thursday was that maybe we can be more consistent about some of the other things we share16:17
cdentthe idea to make it _possible_ for people to know stuff, not to make them know stuff16:17
dmsimardI see the nuance and it makes sense, yes.16:17
cdentfungi: it may be this particular change for some people, but not for me.16:18
dmsimardThe TC summaries and the openstack-dev digests are great at what they do.16:18
fungiand this was mostly my point as well. communication is a shared responsibility between the parties involved. if people aren't receptive to communication then they may be surprised by changes, sure, but that's a choice they make not to get engaged16:18
fungithere is always going to be a trade-off in needing to seek out information to some extent when it's information you want to keep current on16:18
cdentfungi: yes, totally, but _if_ we can do better, we should try. That's all this really is. It's not some condemnation of actions made. It's hope about future actions.16:19
fungii think i agree with you on that point, where we disagree is that i also think we already strike an excellent balance between expending our limited resources in communicating change and using those resources for other important work16:20
fungi"better" isn't a linear equation16:20
fungiit's multivariate, and part of a larger picture of efforts which share a pool of available time16:21
cdentWe may strike that balance, for you, but that doesn't equate with my experience, and I think multiple points of view is part of what makes how we do stuff good.16:22
persiaAlso, "better" is vague without audience definition.  One has to consider who might be surprised.  If one can identify a set of stakeholders with whom communication can be improved, it is easier to target that.  If one is just trying to deal with "someone", it is harder.  "General Public" is a valid target audience, but I have trouble imagining why they would care about internal governance details of the OpenStack foundation.16:22
fungiand also, do we care if some people are surprised by things which they're also not especially interested in learning about?16:23
cdentfungi, persia : Sure, my audience is based on the people I feel like I represent as a TC member. That is, people who talk to me, public or private, about their concerns in OpenStack. Those concerns are either about communication or use of power, so those are the things I focus on.16:24
fungii appreciate that16:24
persiafungi: I hope the answer to that is "false", in part because it is tricky.  Even for fairly obvious things like the value of π, where various organisations are still debating whether people need to be informed or whether discovering it is not a fraction should be surprising.16:24
persiacdent: Declaring your constituency as your target audience is an excellent position to take.  I believe you reach your constituency well with your summary mails, which I would expect anyone who considers you their representative to read.16:25
* fungi wholeheartedly agrees that cdent does an excellent job there16:26
cdentpersia: I hope you're not asserting that everything my target audience might like to know should show up in my emails?16:26
persiacdent: I assert that you should ensure your constituency has sufficient information to be motivated to ensure you remain their representative.  Precisely how you manage to achieve that depends on your particular style as a politician.16:26
smcginnisGet writing. :)16:26
fungiin choosing what to highlight and being somewhat opinionated in your take on those topics, you expose your platform as a tc member, so to at least some extent your constituency is recursively defined by your communication ;)16:27
fungi(i think that goes for all of us, really, in a culture where constituency is defined by what we individually choose to spend time on publicly)16:27
cdentpersia: I was rather hoping that I'd get to the end of this term and be able to say "job done", but I don't.16:28
persiaI am not an active politician today, but when I was, I tended to follow the model of having delegates in each subcommunity upon whose support I drew, and spent time addressing the concerns of the delegates individually each week.  I have seen others do very well with a zero-interaction policy, where they just take a strong position in public, refuse private communications, and ensure their position elicits enough support.  There are lots of valid16:28
persiaways to play politics.16:28
persiacdent: A politician's job is done when either a) there is absolute consensus and no change anticipated in the future, and no constituency seeking change or b) they step aside and let someone else play for a while.  (a) is usually described as "this project has been abandoned" in open source.16:30
*** kumarmn_ has joined #openstack-tc16:30
*** kumarmn has quit IRC16:34
cdentc) their dead16:34
persia(c) is subsumed by (b), except in the case of universal destruction, where it falls under (a).16:37
*** kumarmn_ has quit IRC16:43
fungikeep your eyes on the skies!
*** kumarmn has joined #openstack-tc16:50
smcginnisThe end is near!16:52
dhellmannI've heard a few times recently that stackalytics isn't producing trustworthy numbers. Can we take it down? or find maintainers for it?16:54
mugsieThe last I heard was the current core group did not want to release control?16:55
*** kumarmn_ has joined #openstack-tc16:57
mugsiebut, if there is the opportunity, it is the kind of geeky playing around with data I enjoy so I would be interested16:59
*** kumarmn has quit IRC17:00
clarkbits also run independent of openstack17:01
clarkbso infra can't take it down for example17:01
persiamugsie: If you are sufficiently interesting to stand an official replacement up, at least I would be willing to volunteer to help spread the message that the unofficial one is inaccurate, and your replacement is preferred.17:03
* persia is well aware that mugsie meets any sensible definition of interesting17:03
mugsieclarkb: what would infra need to setup (or something equivelent) ? an infra spec?17:04
mugsiepersia: yeah, interesting is definitley one way people describe me :)17:04
clarkbmugsie: we have a but its not a very happy service17:04
clarkbI think I'm personally not thrilled with the idea of having an official set of stats anyways17:04
persiaI don't like official stats, but I prefer them to unofficial stats that are perceived as official by nearly everyone and regularly mentioned even in board meetings.17:05
mugsieclarkb: getting an NXDOMAIN for :/17:06
clarkboh we must've lceaned it up then due to prolonged unhappyness17:06
clarkbit did exist at one time but like fungi mentioned somewhere recently you have to do a full restart and reindex to get new data17:06
clarkbit also leaks ssh conenctions to gerrit17:06
clarkbstackalytics proper's solution to that problem is to create a new account every time the old one runs out of ssh connections17:07
mugsieI would agree that stats can be problematic - but is (at this point) effectively offical unfortunately17:07
mugsiewait, what? -_-17:07
mugsie /join #openstack-stackalytics17:08
fungiyeah, mrmartin proposed an infra spec (mainly because there wasn't a better means of attempting to communicate intent to the stackalytics devs) to attempt to improve it so it would be a more viable service for us to be able to run reliably17:11
fungi Stackalytics Persistent Cache17:12
fungibut improving stackalytics at this point probably means forking it or trying again to see if they would accept additional developers/reviewers on it17:13
fungioomichi had also expressed interest in hacking on it17:15
fungihe was also interested in gutting most of the local configuration in the repo in favor of queries to the foundation profile lookup api at openstackid-resources.o.o17:16
clarkbiirc red hat expressed interest in it at the boston forum too17:18
mugsiefungi: is that API open?17:18
mugsiethat would be a much better idea than the local config json17:19
fungimugsie: it is, but not terribly well documented17:19
fungiand my attempts to add docs publishing jobs for the openstack-infra/openstackid-resources repo haven't really gotten very far17:20
mugsiebut - we dont't require people to be members to commit code anymore, right?17:20
* fungi checks to see if those changes are still festering17:20
fungimugsie: if you want to extend the member lookup api, you could collaborate with the existing authors/reviewers on right17:21
* mugsie bookmarks17:22
fungiwell, existing author mainly17:22
fungismarcet does hang out in #openstack-infra some but works at a software development firm the foundation contracts to do stuff related to the foundation website (, events schedule/mobile app and related stuff17:23
persiamugsie: I don't beleive we require folk to be members to submit code, but we do strongly encourage it.  Informing organisations that want credit for staff contributions that we expect those staff to be individual foundation members doesn't seem like a bad carrot to me.17:24
fungiunderstanding the data model for one of the underlying databases may involve digging into
mugsiecool - I *think* everything is there (based on the code I see in the repo anyway)17:24
fungi(note that one's not being developed through gerrit)17:24
fungibut yeah, the stuff that stackalytics would want is already queryable anonymously from the lookup api17:25
clarkbpersia: correct though in the past the cla did require it17:25
clarkber not the cla itself. the cla signing process in gerrit17:25
clarkb(was an implementation detail we changed second half of last year)17:26
fungimugsie: correct that we dropped the foundation member-to-code-committer requirement. i think it would be entirely reasonable to tell people that if they want to appear as an affiliated contributor then they should at least set up a profile at to note their affiliation (much like now with stackalytics you have to propose a configuration change to indicate your affiliation)17:27
mugsiefungi: yup, that is true. and the base list of email domains in stackalytics might help as well17:27
fungiwhat's nice is that the member profile data includes things like irc nick, which would allow stackalytics to also track things like irc activity in official channels17:28
fungiso you could have charts showing what times/days people tend to be active in irc, much like you get with the review activity charts now17:29
persiaAt that level of functionality and increase, changing the name to, as suggested above, may have value.17:30
*** jpich has quit IRC17:33
*** andreaf has quit IRC17:39
*** andreaf_ has joined #openstack-tc17:39
jrollI once found an irc stats website, but don't recall at all where that was17:40
jrollI think clark was ranked #1 irc spammer17:40
clarkbI talk too much17:41
cdentcan I exchange irc stats or stackalytics numbers for influencer tokens?17:41
persiaThe eavesdrop data allows one generate reports for arbitrary timespans, and some folk have local logs and may have run against that.  I don't remember a live stats report.17:41
*** andreaf_ is now known as andreaf17:41
persiacdent: Yes.  High IRC traffic almost always generates influencer tokens17:41
jrollquality high traffic does anyway :)17:42
* jroll notes freenode's resident spammers have no tokens17:42
persiaFair.  I assumed that high-volume low-quality traffic was impossible, due to channel access restrictions that are sometimes imposed.17:43
* cdent goes to dinner17:44
cdentHave a good evening.17:44
*** cdent has quit IRC17:45
jrollnight \o17:45
jrollpersia: aha, I found it:
jrollthough looks like it hasn't been updated since january17:56
persiaOh, cool!17:56
jrollsome other neat stats there, no clue how accurate17:56
persia seems to cover some of what was described immediately before17:57
jrollmugsie: maybe this code just needs some updates17:58
smcginnisOh, interesting.17:58
mugsieah, its
jrolloh, cool17:59
smcginnisWas ask.o.o shut down?18:00
smcginnisOh, no. Looks like the stats for that just stopped Jan 2016.18:01
dmsimardjroll: oh wow, I didn't know that existed -- I'll make sure to mention it in the spec I am working on about improving IRC discoverability
mugsiedmsimard: I like that18:14
*** dtantsur is now known as dtantsur|afk18:15
fungiheh, apparently clarkb is the spammiest irc user in our channels in recorded history18:16
fungihe's 22% spammier than me18:16
fungiand that's saying something18:16
clarkbI've fallen off in recent times. There was a time when I tried to lurk all the channels and be helpful at all hours18:16
clarkbthen I had kids18:16
fungiyeah, you're only #6 in the past year18:17
fungiand #9 in the past month18:17
fungia clear trend!18:17
dmsimardfungi: the data hasn't been updated for about a year, the suspense of who is the most spammy contributor itches me :)18:20
jrolloh wow, it's 2018 now, I thought it was just over a month :P18:21
dmsimardI wonder why it just suddenly stopped working, though18:21
*** EmilienM is now known as mimi18:21
*** mimi is now known as EmilienM18:21
smcginnisWhen was the gerrit upgrade?18:22
fungino clue. is basically unmaintained18:23
fungithat was being taken care of under contract with bitgeria but, as i mentioned, we weren't really able to explain our developer and contributor model to them adequately for them to produce useful enough analysis to justify the contract costs18:23
smcginnisOpenStack is always a hard thing to explain to outsiders. :)18:24
*** harlowja has joined #openstack-tc18:24
fungipart of the problem is that they were mostly used to analyzing contribution to software maintained on github or github enterprise18:24
fungiso they kept trying to relate our methods back to "equivalent" parts of their github-oriented data model which weren't really equivalent at all, so as to avoid needing to alter their software and assumptions too much18:25
fungiwe kept trying to point them back to query apis we have to get the relevant data, but they continued to do things like raw analysis of git commits in repositories (which ends up double-counting contributions to forked repositories, confused by the predominance of merge commits due to our highly parallel workflows, counting named references gerrit replicates for each patchset of every change whether it18:28
fungiwas ever merged to a branch or not...)18:28
fungias a result the numbers they came up with were not relevant18:29
fungistackalytics did/does a better job of modelling how we work, for obvious reasons18:31
dmsimardInteresting -- I would tend to think that they would be interested in implementing Gerrit in general, seeing as there are a lot of big projects (Android comes to mind) using it18:34
smcginnisBut how can you rank a project against others if you don't have all those github stars?18:35
dmsimardOh, it does mention that gerrit is supported here: (whatever that means)18:35
dmsimardsmcginnis: do we really want to talk about that cncf website again? :p18:35
smcginnisHaha, nope!18:35
*** kumarmn has joined #openstack-tc18:38
*** kumarmn_ has quit IRC18:42
*** rosmaita has quit IRC18:43
fungiit's also been a couple years since we engaged them, so possible they've improved their support for querying gerrit in that timeframe18:44
fungibut the other upshot was that the foundation had been paying them primarily to have quarterly and annual reports to provide to the board on project activity, and over time determined that the cost of paying a third party to do that wasn't worth the value we were getting compared to just whipping up some stats ourselves for things we wanted to highlight in those reports (also the board didn't need/want18:46
fungianywhere near the level of detail we had been providing)18:46
fungithe activity dashboard was really just a byproduct of that18:47
smcginnisThat makes sense. I would hope a lot of the really needed data points would be easy enough for us to do on our own without needing to write out large checks to third parties.18:48
pabelangerheh, organization for seems incorrect for users19:30
pabelangerat least for me19:30
*** kumarmn has quit IRC19:31
smcginnisThere's a lot missing.19:31
smcginnisI tried to check out the cinder activity but nothing comes up, so I don't think it actually has all data.19:31
pabelangeryah, I forgot about that site. Don't really know how it works19:31
smcginnisAnd we did notice it doesn't appear to have updated since January 2016, so those organizations may have been correct for the time.19:32
pabelangerseems to show last sync19:33
pabelangerI'm not sure how manages that TBH19:34
*** annabelleB has quit IRC19:58
*** annabelleB has joined #openstack-tc20:10
fungipabelanger: nobody now. we basically provided an unconfigured server instance to the devs from bitgeria and they set the server up and managed it. given it has only outdated data for over a year, it's probably well past time to discuss taking it down20:19
fungiwhere "we" is the foundation admins, not the openstack infra team20:21
fungiit's not infra team control at all20:21
fungier, controlled20:21
pabelanger++ if unmanaged, taking it down seems right20:23
*** rosmaita has joined #openstack-tc20:56
diablo_rojofungi, we can definitely talk about sponsor involvement in upstream development at the fc-sig meeting. Not sure what sort of power we have or question you want us to answer though.21:32
*** flwang1 has quit IRC21:40
fungidiablo_rojo: more like helping find a good place to keep a list of what sorts of technical resources contributors will need (e.g., access to irc, ability to send e-mail to/receive from mailing lists, access to gerrit, and so on)21:44
diablo_rojofungi, you just described the contributor guide21:44
fungisure, except communicated in a way that they can pass along to their corporate management, mailserver operators, firewall admins and so on21:45
diablo_rojoI guess its not a specific list of things you need so much as directions as to how to configure them, but basically there.21:45
diablo_rojoOkay so something less how to and more requirements21:45
fungiwell, less a list of things you as a new contributor need to be familiar with and more a list of things your corporate overlords need to possibly make allowances for you to have access to21:46
diablo_rojoGot it. We can make that list in the FC SIG wiki for now? And maybe find space in the contributor guide later? Make it a part of common or something.21:46
diablo_rojoHas anyone started compiling this list by chance?21:47
fungithe conversation started pointing out that we have contributors coming from companies who won't let them access irc or send e-mail to outside addresses from work21:47
fungiand trying to figure out how we get the word out that these are things they will need to provide their employees for them to be successful participants in the community21:47
diablo_rojo'What your company needs to allow and provide for you to succeed in OpenStack'21:48
*** flwang1 has joined #openstack-tc21:53
fungisomething along those lines, right22:17
fungiwe have lots (and an increasing amount) of documentation helping new contributors learn what they need to succeed, but nothing i know of targeted at the companies they work for letting them know what to expect/provide to help their employees succeed in our community22:18
fungiwhich was a big part of the original remit of the fc sig... helping provide guidance to contributing companies and not just individual contributors22:19
fungi(though the two certainly overlap a ton)22:19
persiaMy general thought in this area is that the content is nearly 100% overlap: it's just the tone that differs.  I suspect that the details of "how do I" are best as links to the individual contributor guide (so skipped), and it becomes more "What do my staff need to be productive?" (firewalls, off-hours access to network/email, etc.), "How do I influence priorities for a project?" (travel budgets, planning cycle deadlines, etc.), "What sort of22:35
persiadonations from my organisation are credited?" (developers engaged in reviews/commits, cloud resources for infra, etc.)22:35
clarkbI would be careful about off hours email access and pushing that. Often that implies people have to be paid outside of their normal work hours22:37
persiaYes.  I assert that is critical to success in OpenStack.  If it is true that a given meeting always happens during regular work hours for someone, then it is true that it never happens during regular work hours for someone else.  Either we share that burden in some way, or we forego realtime collaboration.22:40
persiaThe key is off-hours access to network (although email is often helpful also, especially if referring to something for IRC).22:41
clarkbI agree, but I also think it isn't unreasonable for individuals to have lives and not be required to work crazy hours. ANd making that a hard requirement may not be realistic for all contributors22:42
persiaThere are alternative ways to deal with timezones, but the openstack norm seems to be for people to work odd hours.22:42
persiaMost of the folk with whom I have spoken are willing to deal with meetings from 6am to midnight, and most are able to get employer acknowledgement of odd times in exchange for less time in the office.  At PTG, there were a few folk who had been attending 2:00 meetings, and were happy to be able to discuss moving the meeting with a PTL (as it is difficult to do both a 2:00 meeting *AND* get the kids out the door at 7:30).22:43
persiaYes, this is madness, but, as I said, the alternative is to not do meetings at scheduled times, or ignore large portions of the planet.22:44
persia(one of the difficulties with being an APAC contributor is that often folk don't do the timezone math carefully, so "APAC friendly" meetings sometimes happen at fairly inconvenient times)22:44
persia(or so I gather from speaking with folk who live in large positive timezones in Dublin)22:45
diablo_rojoI feel like we should start an etherpad that we can track needs in and how we want to phrase them22:45
persiadiablo_rojo: Excellent idea.22:45
diablo_rojoand after a time, add that to the contributor guide22:45
diablo_rojoAny brilliant ideas for a name?22:46
persiaContributing Organisation Guide22:46
persiaOr maybe with oxford 'z' :)22:46
diablo_rojoOrganization like a company?22:46
diablo_rojogot it22:47
persiaI just am lazy about my orthography.22:47
diablo_rojoclarkb, fungi persia feel free to add to it as you see fit ^^22:49
*** hongbin has quit IRC22:50
diablo_rojoAt least document the struggles you have heard of and we can try to form it into a requirement/need22:50
clarkbpersia: at least in the infra team we've tried to be flexible around the meeting and do our best to record important notes with the status bot logging. Beacuse ya single meeting for global team is hard and at least one timezone is likely to be in a bad situation22:50
persiaclarkb: And I watch people attend the infra meeting at 6:00 and 20:00 every week.22:51
clarkbyup some do22:51
clarkbdst makes it easier and harder at different time of the year too22:52
clarkbit is especially difficult for southern hemisphere because the time delta relative to northern is 2 hours and in the wrong direction once a year22:52
persiaYes.  UTC+10 to UTC-7 is easier than UTC+11 to UTC-8.22:53
persiaI think soon the infra meeting attendees will be attending from 5:00 to 21:00, but maybe some will drop off.22:53
persia(or maybe I am failing at timezone math)22:54
clarkbreltiave to utc its only an hour so for the meetings its not as terrible22:55
clarkbbut just general being awake during overlapping hours22:55
persiaMy point being that because nearly every team has meetings like that (some are scheduled to be very awkward from the Americas), I think it unwise to assert that it one can be as effective in OpenStack if one doesn7t have remote access to comms.22:55
cmurphywe should not be a community that requires people to work odd hours in order to participate, we should be leaning more on the mailing list and less on meetings22:56
persiacmurphy: Changing our culture is a fine thing.  While I like realtime cultures (and have adapted my life around them), there isn't any reason not to try to be more inclusive.  I only describe the culture that exists, and which causes you to be involved at midnight.22:57
cmurphyI have few enough obligations in my life that I can afford to be involved at midnight, that's not so for everyone and I would rather push toward a culture where we're not promoting working at all hours of the day as the norm23:01
persiaA first step towards that would be to abolish timed meetings.23:02
persiaFor the first few years of my OpenStack involvement, I was not an Infra contributor (despite this being my prime area of interest) mostly because I had trouble making a 4:00 meeting every week, given other responsibilities I had at the time.  I don't mean to pick on a single team, but I have firsthand experience with the difficulties of working in an OpenStack project in the wrong timezone.23:04
persiaeven if we just restrict to folk living between UTC+1 and UTC-8 (or UTC+2 and UTC-7 at certain times of year), many meetings fall outside a normal workday.23:06
clarkbits definitely not perfect23:08
clarkbI do think it is something we have improved on slowly though23:08
persiaI'd be curious how we've done that.  From discussions at PTG, I had the impression that folk in UTC+8 and UTC+9 either did crazy things or just ignored the westerners, and folk in UTC+10 through UTC+12 just got up early, had lazy mondays, and worked saturdays.23:12
persiaThat might be an improvement from before, but I think it is worth tracking what specific changes have addressed timezone concerns.23:13
clarkbpersia: one example is tc siwtching from single meeting a week to office hours that span timezone friendly blocks of time23:13
clarkbanother is there is definitely more of a reliance on specs and the mailing list for sweeping changes23:14
persiaDid I still live in UTC+9, I would have a harder time attending the "APAC friendly" TC office hours than I do now.23:14
clarkbspecs in particular I think havehelped communicate far better planned action than blueprints in lp23:15
persiaAlso, few TC folk show up most weeks.23:15
persiaI agree specs are better than blueprints.  I'm unsure how that affects timezones.23:15
clarkbpersia: because you no longer need to be in the meeting each week or on irc during "peak" hours to understand large scale changes to projects23:16
clarkbits still not going to be perfect but I do think it was a great improvment over the situation before23:16
persiaGiven the current thread about how implementations and specs don't match, I'm uncertain, but maybe it is a bit better.23:16
jrollwhile I don't think we should promote working odd hours, and I also think we should work to make that unnecessary - I do agree that we should have something in this guide about upstream developers having access to the resources they need at all hours, in case they make the choice that they do want to participate outside of normal working hours (as doing so can help one get their work done).23:17
persia(and I have a hard time arguing from the "I'm timezone challenged" perspective, because I happen to have a schedule in a fictional timezone that matches well with just about everyone's workday (ignoring the same folk everyone else ignores))23:17
clarkbjroll: yup if people want to be working at midnight (I've been there it was actually preferable) they shoudl eb able to23:18
jrollfor example, when I lived in california, I would regularly start my day at 0600 to be able to have conversations with some of ironic's EU contributors. not because it was required, but it made my life easier23:18
jrollclarkb: cool, I just want to make sure we don't block reality because it isn't ideal :)23:19
persiaI've tried to capture the nuance in the etherpad.  Please correct if you feel I've missed an important point.23:20
clarkbpersia: ya I think that captures it well23:22
* jroll adds one note23:23

Generated by 2.15.3 by Marius Gedminas - find it at!