16:00:16 #startmeeting api-wg 16:00:26 Meeting started Thu Aug 3 16:00:16 2017 UTC and is due to finish in 60 minutes. The chair is cdent. Information about MeetBot at http://wiki.debian.org/MeetBot. 16:00:27 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 16:00:29 The meeting name has been set to 'api_wg' 16:00:31 \o 16:00:32 o/ 16:00:33 hi 16:00:40 #chair edleafe elmiko 16:00:41 Current chairs: cdent edleafe elmiko 16:00:48 #link agenda https://wiki.openstack.org/wiki/Meetings/API-WG#Agenda 16:01:07 #topic old biz 16:01:13 #link old minutes http://eavesdrop.openstack.org/meetings/api_wg/2017/api_wg.2017-07-27-16.00.html 16:01:37 elmiko: has the action to create a thing and he did! https://review.openstack.org/#/c/487847/ 16:02:04 * edleafe pats elmiko on the back 16:02:04 i missed dtantsur's comments by about 1 minute when i pushed rev 4, i'm gonna fold them into the next rev 16:02:06 and we saw it and said it was good 16:02:20 cool 16:02:21 but, i think he raises a good question about storing the results 16:02:44 oh yeah, publishing is a great point 16:02:49 dtantsur++ 16:02:59 :) 16:03:04 how should we approach that though? (email, wiki, etc) 16:03:22 i'm just thinking what should go in the review process doc 16:03:28 email or the api-wg repo, I would say 16:03:41 in the repo I would think 16:04:22 is that the guidance repo, or? 16:04:37 perhapd as clarifications/examples on existing guidelines? or maybe simply as a separate collection of concrete examples with links? 16:04:42 I think in the guidance somewhere is right 16:04:46 ok, cool 16:04:55 i kinda lean towards having them as separate docs 16:05:23 we may link to them from guidelines as examples, when appropriate 16:05:26 i feel like maintaining the linkage between reviews and guidelines might become cumbersome 16:05:41 for sure, as long as it doesn't incur a crazy amount of tech debt 16:06:08 So sort of a "war stories" document? 16:06:15 yeah, i think so 16:06:37 the cross-linkage thing, while it would be awesome, i just worry about the long tail 16:06:59 yeah, war stories sounds right/good 16:07:12 and less effortful than integrated with existing guidelines 16:07:24 if we make it too effortful, nothing will happen 16:07:38 * cdent looks at everyone over his glasses 16:08:02 hehe, agreed 16:08:08 dtantsur: is "war stories" understandable? 16:08:16 I don’t think we should use that term 16:08:30 no no, something more like "case studies" 16:08:33 well not in the doc 16:08:45 I meant for the purposes of this discussion 16:08:50 oh yeah 16:10:19 edleafe: tbe first time I hear this term, but don't rely on my English :) 16:10:21 ok, so i'll put up some changes after the meeting incorporating dtantsur's nits and what we are talking about 16:10:22 "Case studies" sounds too formal. Maybe just "discussion archive" 16:10:47 +1, i'll go more that direction edleafe 16:11:00 "Stone tablets of inevitable awesomeness" 16:11:04 haha 16:11:05 dtantsur: as soon as I wrote it I thought it may not translate well 16:11:15 we’ll figure it out, and if we find we no likie, we can change 16:11:21 dtantsur: but I like your suggestion! 16:12:17 #action elmiko to continue with review process doc 16:12:25 #topic new biz 16:12:40 #link Do we become a SIG? http://lists.openstack.org/pipermail/openstack-sigs/2017-July/000022.html 16:12:48 are people aware of that thread? 16:13:04 * dtantsur is 16:13:10 me too 16:13:19 I think we're a good example of a SIG, actually 16:13:21 i read some of it 16:13:37 yeah, i don't see any reason not to become a sig 16:13:51 dtantsur: me too, until ttx said moving our mail to the -sig list. At that point I thought “hmm, there goes our main audience” 16:13:52 but, i admit i might be missing some nuanced level of this process 16:14:01 mmmmmm 16:14:03 like that detail 16:14:05 it does seem like for us it would be not much more than a name change 16:14:18 I missed this bit. Why can't we use [tags] like everyone? 16:14:22 I don’t know that there is much nuance. It is just a shuffling of pieces on the board to try to draw more attention to the need for boundary blurring 16:14:37 ahh, gotcha 16:14:38 cdent: we could move internal stuff to -sig, but weekly updates should still go to -dev 16:14:58 (as if we have much internal email discussion) 16:15:00 do we have much internal stuff? 16:15:02 aja 16:15:08 * aha 16:15:11 edleafe: but then we’re not doing the desired boundary blurring. As the intention is to reach a broader audience 16:16:29 cdent: separating email communication hardly seems like boundary blurring 16:16:30 but it seems like our primary audience is openstack devs 16:16:46 in specific, openstack project devs 16:16:56 edleafe: what I mean is that if we only send the newsletter to -dev, we are not reaching ops/users/pms/etc 16:17:24 moving anything to another mailing list will complicate things like "[api] [ironic] What to do about your API, folks?" :) 16:17:25 the point of sigs to make it easier for everyone to be encourage to become a contributor in many various ways 16:17:36 I mean, several tags, including us and not us 16:17:37 yeah, I agree it is messy 16:17:48 dtantsur: ++ 16:17:56 which is why I brought it up 16:18:43 at this point, i feel i have to ask are we looking to solicit more input from those other categories ops/users/pms ? 16:18:58 not trying to say we reject them, but i thought we served the dev community mostly 16:19:06 we want for sure 16:19:22 cdent: we can send the newsletter to whatever lists we want today. Changing to a SIG doesn't affect that 16:19:35 I think we do need input from those communities because without it we have less insiight into the inconsistencies 16:19:37 if you followed (cdent did) the discussion re api interop tag, one of my complains was being dev-centric 16:19:50 cdent: and i agree with that sentiment 16:19:57 * openstack dev centric (/me cannot English after 6pm) 16:20:01 edleafe: yes, I’m saying that maybe that should change if we buy into sig-ness 16:20:16 dtantsur: a fair point 16:20:56 cdent: to my thinking, we should do that whether we are a WG or a SIG or whatever 16:20:56 maybe i'm just skeptical about the involvement of ops/users/pms 16:21:14 edleafe: yes, that’s the underlying point 16:21:36 elmiko: that’s fair, because there’s not that much involvement from devs either ;) 16:21:38 cdent: so remind me again what this has to do with becoming a SIG 16:21:47 cdent: right! 16:22:45 So the reminder is this: As there is this move on to become a SIG, shall we use that as an excuse to expand our audience and publish output and solicit input more widely, or are we in fact an openstack-dev oriented org and as such we should not become a SIG? 16:22:53 Does that encapsulate the issue(s) better? 16:23:06 this is an interesting way to put it, hmm 16:23:21 I'm all for listening to people we seem to help :) 16:23:23 cdent: that sort of begs the question 16:23:41 it is for the begging of the question that we speak 16:23:47 of course we want to be as broadly relevant and helpful as possible 16:23:48 cdent: it does encapsulate well, and at the same time makes the question much more interesting to digest 16:24:00 edleafe: +1 16:24:31 edleafe: is that an of course? I think some of the time we think so, but earlier in the conversation we had comments to suggest otherwise. 16:25:20 i can totally understand the angle of accepting more input from outside contributors, that makes sense. but what are the long term goals from interacting with a wider community? for example, would we see a day where the api-wg is providing api advice to projects that merely built on top of openstack? 16:25:21 _I_ think it is an of course, but I also fear that we may not have the resource of wherewithal to be useful more broadly (not that I can really enumerate what the broader obligations would be). 16:25:28 jinxish 16:25:38 cdent: what sort of comments were they? 16:25:40 heh 16:25:46 * edleafe has swiss cheese memory 16:26:06 edleafe: elmiko’s comments “but i thought we served the dev community mostly" 16:26:14 elmiko: if we're not confident in our guidelines enough to recommend them to people outside of the TC coverage, then something is wrong 16:26:48 dtantsur: oh, i think we are, it's just a question of outreach. will we be directly interacting with folks from outside the openstack dev community in the same capacity? 16:27:15 if they come to us, why not? there is a whole trend on getting closer to e.g. container communities 16:27:31 like, our guidelines are useful, they provide a nice conversation point for starting design/analysis. but they are mainly focused on the works of the openstack community 16:27:33 cross-community outreach is on trend 16:27:34 cdent: we may have arisen from the dev community in response to shortcomings in the dev community, but that doesn't mean that's all we should be 16:27:38 I've heard some folks can learn from our API versioning (no irony this time) 16:27:52 edleafe: I agree, you don’t need to convince me. 16:28:20 dtantsur: that topic will come up shortly when we get to guidelines 16:28:45 i guess i now have mixed feelings about the sig expansion 16:29:05 i'm a little nervous about the scope expansion of the group given our size and history 16:29:10 cdent: but it's no fun if we don't argue 16:29:35 elmiko: there's always the chance that we might attract new blood 16:29:53 edleafe: definitely, and that could be good or bad depending on the situation 16:30:17 like, i sit in on some of the sig meetings for kubernetes and i just wonder how much additional density we could handle 16:30:45 like, if our group got popular in a cross-project way, it could become quite hectic in a short amount of time 16:30:58 not saying that's a bad thing, just a risk to understand 16:31:34 how much thrashing could we withstand kind of thing 16:32:08 * elmiko steps off soapbox 16:32:09 better to burn out than fade away 16:32:21 or something 16:32:22 REST never sleeps 16:32:29 haha, i dunno, i've been thinking about that phrase in my personal life recently and i think i prefer fade away 16:32:40 heh, me too 16:32:44 LOL 16:33:10 :D 16:33:42 so, maybe we should consider it for a little more time then hold a vote or something at an upcoming meeting? 16:33:54 i'm not sure the best way to proceed 16:34:17 I don’t think there’s any rush. Pausing for reflection is warranted and fine. 16:34:31 ok, cool. i didn't know if we needed to decide by the ptg or something 16:34:35 no 16:35:17 we should, however, reflect soe of our concerns back to that email thread, so that the notion of us being an early adopter isn’t as simple as “yeah, sure” 16:35:27 +1 16:35:53 I can do that if nobody else wants to, but it won’t be today or tomorrow 16:36:25 I was sort of hoping to have a SIG discussion at PTG 16:36:46 I'm still not clear on how it would affect us long term 16:36:58 edleafe: +1 i'd like that too 16:37:13 you mean a discussion about sigs? 16:37:41 if so, I’d suggest saying so in response to ttx’s thread (the link above): More info required please. 16:37:53 CLOSED INSUFFICIENT_DATA 16:37:55 yes, i do mean that 16:38:04 and yeah, i'll make a response 16:38:22 i'm not sure the specifics as i've been getting digest emails but i'll figure it out (any advice appreciated) 16:39:07 any more on that or or any other new biz/open topic? 16:39:31 #action edleafe to respond to ttx on ML 16:39:38 #undo 16:39:39 Removing item from minutes: #action edleafe to respond to ttx on ML 16:39:45 #action edleafe to respond to ttx on ML regarding SIGs 16:39:52 +1, thanks edleafe 16:40:18 okay 16:40:24 #topic guidelines 16:40:29 #link https://review.openstack.org/#/q/status:open+project:openstack/api-wg,n,z 16:41:28 the thing I wanted to mention about versioning was that sdague has had a talk accepted at apistrat about microversions. As part of that he’s going to be updating the doc at https://review.openstack.org/#/c/444892/ more vigorously in prep 16:41:57 ooh, neat! and congrats to sdague 16:42:16 the latest rev of that doc wasn't changed much 16:42:34 yeah, I’m very curious to hear how people respond. There will be some purists who will react very poorly, and some pragmatists who will be interested 16:42:52 cdent: are you gonna be there as well? 16:43:04 not last I checed 16:43:11 ack 16:43:16 but I’m hoping to hear about it on the grapevine 16:43:23 yeah for sure, was just curious 16:43:59 it’s the sort of thing I’d like to go to, but not even thought of investigating until about 2 minutes ago 16:44:19 haha 16:44:37 we’ve already talked about the review process guideline 16:44:39 so that leaves 16:44:45 #link https://review.openstack.org/#/c/487504/ 16:45:00 does that count as a non-guidance change? 16:45:22 * elmiko looks 16:45:35 it’s basically repointing links 16:45:39 Yes, IMO 16:45:48 yeah, i think so too 16:46:05 Shall I pull the trigger? 16:46:33 go fer it 16:46:46 Done 16:46:47 \o/ 16:46:51 #topic bug review 16:46:58 when that merges we’ll have one less bug 16:47:03 #link https://bugs.launchpad.net/openstack-api-wg 16:47:23 Hey, I had a question about the Monty series 16:47:39 I tried rebasing it, and got stuck on the middle patch 16:48:04 Can I push it with the one conflict I couldn't resolve? 16:48:34 Of course, I'd -1 it until Monty can fix it 16:48:41 good with me 16:48:44 that sounds fair 16:48:48 ok 16:51:14 when (in the mythical future) placement allocation handling resolve, I ought to be able to pick up one more buglet from the bug list. I’m going to try to do that regularly, but I’m not sure when the mythical future gets here. 16:51:32 +1 16:52:21 anything else on bugs? 16:52:53 nope 16:52:55 #topic weekly newsletter 16:53:00 #link https://etherpad.openstack.org/p/api-wg-newsletter 16:53:13 I can’t do it this week, I need to run off an fetch my wayward lovely wife from the train station 16:53:35 * elmiko hides behind the couch 16:53:40 Guess it's my turn, huh? 16:53:52 either that or we flip a coin 16:54:02 Heh, as if I had any money 16:54:04 i have time to do it 16:54:06 lol 16:54:08 I'll do it 16:54:11 thanks 16:54:18 thanks edleafe 16:54:26 anything else from anyone about anything? 16:54:51 * edleafe has nothing 16:55:06 * dtantsur too 16:55:23 nothing from me 16:55:34 although, anything you say....? 16:55:37 XD 16:56:36 So are we done? I'll ping you in -sdks when the newsletter is done 16:56:44 we done 16:56:48 thanks everybody 16:56:52 #endmeeting