16:00:20 <cdent> #startmeeting api-wg 16:00:23 <openstack> Meeting started Thu Apr 7 16:00:20 2016 UTC and is due to finish in 60 minutes. The chair is cdent. Information about MeetBot at http://wiki.debian.org/MeetBot. 16:00:24 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 16:00:27 <openstack> The meeting name has been set to 'api_wg' 16:00:30 <elmiko> hi 16:00:33 <etoews> o/ 16:00:36 <cdent> hola 16:01:00 <cdent> #link https://wiki.openstack.org/wiki/Meetings/API-WG#Agenda 16:01:01 * elmiko does 3 amigos salute 16:01:15 <cdent> anybody else joining in for api-wg meeting? 16:01:35 <cdent> #topic previous meeting action items 16:01:35 <cdent> #link http://eavesdrop.openstack.org/meetings/api_wg/2016/api_wg.2016-03-24-16.00.html 16:01:49 <dstanek> o/ 16:02:11 <cdent> I reckon those are done. I made https://launchpad.net/openstack-api-wg and started https://review.openstack.org/#/c/301846/ 16:02:47 <cdent> and etoews even made a bug: https://bugs.launchpad.net/openstack-api-wg/+bug/1562058 (gaining him ten points!) 16:02:48 <openstack> Launchpad bug 1562058 in openstack-api-wg "Links guideline" [Undecided,New] 16:02:52 <elmiko> nice 16:03:02 * etoews loves internet points 16:03:11 <cdent> you can do so much with them 16:03:31 <cdent> okay, on to new biz 16:03:37 <cdent> #topic etags and the lost update problem 16:03:57 <cdent> If you haven't already seen it, that guideline has gone quite long 16:03:58 <cdent> #link https://review.openstack.org/#/c/301846/ 16:04:05 <cdent> and I have some questions/concerns about it 16:04:22 <elmiko> i thought the baldly thing was a joke at first XD 16:04:37 <cdent> Though long it leaves out a lot of detail 16:04:58 <cdent> baldly as in: overly straightforward, without even being shy about it 16:05:36 <elmiko> agreed about the length, but it seems like a great intro to someone who doesn't intimately understand etags 16:05:42 <cdent> I had friend (from a similar we-love-http background) gaze over it and his first reaction was akin to "this is way dumbed down" 16:05:57 <elmiko> huh 16:06:11 <cdent> "why don't you talk about state more?" 16:06:18 <elmiko> but as a guideline, how far do we drill down before recommending outside sources (which you did) 16:06:24 <cdent> "shouldn't the term 'conflict' show up more often?" 16:06:33 <elmiko> ah, fair 16:06:34 <cdent> yeah, exactly 16:06:39 * etoews reading 16:07:16 <elmiko> and, imo, the state talk starts to get into impl side of things 16:08:04 <cdent> On the flip side the prose pretty much just flowed out of me, so we can probably get me to write more like these if we think this form of document is a good idea (that is, the patented elmiko-driven notino of problem->solution->details) 16:08:45 <elmiko> i'm glad to hear it flowed so well, i certainly thought it read well. i'm curious to hear if we have missed anything big with this style. 16:10:01 <cdent> Do any existing openstack services already use etags? I think we decided "no". 16:10:07 <elmiko> swift? 16:10:34 <cdent> does swift count? :) 16:10:48 <elmiko> your words, not mine XD 16:11:14 <dstanek> cdent: i just did a quick read and it looks like a good intro 16:11:17 <elmiko> but, aside from that, i think no 16:11:31 <cdent> thanks dstanek 16:11:38 <elmiko> dstanek: that's encouraging to hear 16:12:29 <cdent> etoews: still reading? 16:12:44 * etoews making 1 comment 16:13:06 <cdent> rad 16:15:50 <cdent> Are there additional "new topics" we should be talking about (stuff that won't be covered by guidelines and apiimpact further down the agenda)? 16:16:09 <etoews> elmiko: are we going to prepare anything for the summit session? 16:16:11 <elmiko> do we need to discuss the summit session any further? 16:16:14 <elmiko> hehe 16:16:16 <etoews> :) 16:16:49 <elmiko> i have a couple proposals to bring to the group around PR for the guidelines, but i'd need at most like 15 minutes to present 16:16:58 <etoews> go for it 16:17:33 <dstanek> is there an etherpad for summit session content? 16:17:48 <elmiko> https://etherpad.openstack.org/p/austin-api-wg-session-plans 16:17:58 <elmiko> i think we need to clean it up a little though 16:18:10 <cdent> etoews: you wanna talk about your comment on etags here or should we save it for the review? 16:18:18 <etoews> elmiko: after you present we just forge out way through the session agenda? 16:18:20 <dstanek> elmiko: perfect, thanks 16:18:42 <etoews> cdent: save it for the review. i'm going to have to bounce at any minute. 16:18:49 <cdent> cool 16:19:03 <cdent> I've got one more quick "new topic" that I forgot to put on the agenda 16:19:12 <cdent> #topic tc candidates love api-wg 16:19:31 <etoews> :D 16:19:35 <cdent> I didn't make a close count but a lot of the candidates for TC mentioned the api-wg and how it is going to take a big role in the future 16:19:46 <elmiko> etoews: i'm making a few comments on the pad, but i see; state of the wg, upcoming activites, and promiting the guidelines 16:19:57 <elmiko> sweet! 16:20:07 <cdent> that's either a cross to bear or an opportunity, but either way I guess we need to be aware of it 16:20:32 <etoews> hah. agreed. 16:20:41 <elmiko> also, not directly api-wg related, but since we share a channel i absolutely love the work that is happening with the openstacksdk. big props to those folks 16:20:42 <etoews> i gotta run. 16:20:49 <etoews> see you later in irc 16:20:53 <elmiko> later 16:20:57 * cdent waves to etoews 16:21:39 <cdent> Yeah, I think one of the side effects of the big tent is that things that might have once have been thought of as on the edges are now become more central 16:21:52 <cdent> because they links stuff up 16:22:35 <cdent> But yeah, just wanted to air that observation. 16:22:39 <cdent> Shall we move on? 16:22:56 <elmiko> was sdk seen as on the edges before? 16:23:09 <elmiko> or was that a comment about api-wg 16:23:10 <cdent> Well one way to put it would be "it's not nova" 16:23:15 <elmiko> haha! 16:23:33 <cdent> and somewhat more politick: it's not a service 16:23:36 <elmiko> so, this is like galileo declaring that openstack does not revolve around nova? 16:23:44 <cdent> and the services tend to accumulate the gravity 16:23:50 <cdent> something like that 16:24:14 <elmiko> yea, definitely cool to see more projects growing 16:24:19 <cdent> or maybe gravity trumps planets 16:24:28 <elmiko> heh 16:25:03 <cdent> #topic guidelines 16:25:12 <cdent> #link https://review.openstack.org/#/q/status:open+project:openstack/api-wg,n,z 16:25:36 <elmiko> hmm, when are we merging all those frozen ones? 16:25:47 <cdent> probably now 16:26:06 <elmiko> we should merge this https://review.openstack.org/#/c/299416/ 16:26:34 <elmiko> not a guideline, so i went for the +2 16:27:46 <elmiko> ok, so aside from the redundancy one, you should merge the microversions and unexpected attr one (cdent) 16:28:16 <elmiko> we did announce those on list right? 16:29:28 <cdent> the guidelines were clarified such that we only go to list when we have contention 16:29:57 <elmiko> ah, ok. i thought we still announce when they freeze (which i think you did with those) 16:29:57 <cdent> because the intro to the guidelines suggested we were supposed to be prioritizing velocity and forgiveness 16:30:15 <cdent> yeah, I still like the idea of emailing frequently 16:30:20 <cdent> and usually do anyway 16:30:32 <cdent> I've +w those 4 16:30:32 <elmiko> right, ok 16:30:41 <elmiko> sweet, cdent++ 16:30:44 <cdent> so we can send an announcement for them 16:30:49 <cdent> shall I, or you wanna? 16:31:04 <elmiko> i thought we only email when we freeze, not merge? 16:31:42 <cdent> when I was teasing apart the guidelines (which were way confusing) 16:31:45 <cdent> where it ended up was: 16:31:57 <cdent> "An email should be sent to the openstack-dev mailing list containing the links to all of the guidelines that have recently merged. The finalized guidelines should be buffered such that a maximum of one announcement email is sent per week." 16:32:04 <cdent> which is unchanged from the original 16:32:13 <elmiko> ahh, ok 16:32:28 <elmiko> we may have been lax on that, but yes i will send an email 16:32:36 <cdent> mind you, I agree with you that that's a bit weird 16:32:41 <cdent> I would rather have this order: 16:32:42 <cdent> we review 16:32:44 <cdent> cpls review 16:32:53 <cdent> opened for wide review by email announcment 16:32:56 <cdent> merge 16:33:08 <elmiko> i /think/ that is what we have been doing 16:33:09 <cdent> however I think the concern was that this would be slow and invite input from crazies 16:34:42 <elmiko> maybe we should just ammend the guidelines to remove the post-merge email 16:34:55 <elmiko> unless we decide to move to some sort of "api-wg status" type thing 16:35:33 <cdent> perhaps at summit you can talk with the people there about the feelings towards email and the risk of crazies 16:35:44 <elmiko> ok 16:35:46 <cdent> I think the guidelines were created by etoews and cyeoh 16:36:03 <elmiko> i don't think it's been an issue though, we have always sent out an [all] email on the frozen guidelines 16:36:10 <cdent> I think I've made my position pretty clear above? "more input from everybody, yay!" 16:36:36 <elmiko> yea, and i don't see a deviation from our current proceedure 16:36:39 <cdent> #action at summit elmiko will engage the people(tm) to see about how and when email 16:36:58 <elmiko> i'll add an agenda item to discuss the guideline process again 16:37:06 <cdent> cool 16:37:35 <cdent> some of the older guidelines seem a bit stalled 16:37:58 <elmiko> yea 16:38:11 <elmiko> i've been meaning to get back to the actions guideline, but i feel it's a massive can of worms 16:38:26 <elmiko> and in need to send an email to help distill the thoughts 16:38:56 <cdent> yeah, it's difficult 16:39:35 <elmiko> maybe we should codify something about stale guidelines? 16:39:43 <elmiko> (i'll add it as a topic for summit) 16:39:54 <cdent> I'm hesitant to become too overloaded with process 16:40:04 <elmiko> good point 16:40:23 <elmiko> i'll add something about creating a process to limit the process..... ;P 16:40:49 <cdent> hehehe 16:40:53 <elmiko> maybe just something to be aware of then (stale guidelines) 16:41:23 <cdent> review handling/velocity in general is a bit of a chore, especially things with apiimpact (which I almost never look at :( ) 16:41:37 <elmiko> yea, that is very difficult 16:42:03 <elmiko> i've tried to review the PRs that i've been pulled into. but it's tough to get them all 16:42:18 <elmiko> i think we need to lean more on having good CPLs 16:42:35 <cdent> yes 16:42:38 <elmiko> people who can help determine when a project needs more outside assistance for api stuffs 16:42:43 <cdent> other than dstanek there's nobody here 16:42:48 <elmiko> right 16:42:49 <cdent> we probably need to send reminder emails 16:42:58 <elmiko> and don't get me started about the 0000UTC meeting... 16:43:07 <elmiko> imo, we need to go back to weekly meetings 16:43:11 <cdent> yes 16:43:24 <cdent> another summit topic? 16:43:26 <elmiko> biweekly just kills the cadence 16:43:27 <cdent> I can do whenever 16:43:29 <elmiko> yea, i'll ad it 16:43:29 <cdent> yes 16:43:48 <cdent> quiet revolution 16:44:05 <elmiko> lol 16:44:30 <elmiko> i mean, in the beginning i had a few nice private chats with cyeoh in 0000 slot 16:44:39 <elmiko> but now it's just empty 16:45:02 * cdent raises a glass 16:45:11 * elmiko raises a glass 16:45:17 <dstanek> cdent: :-) 16:45:37 <elmiko> added a topic for summit about CPL participation as well 16:45:46 <cdent> excellent 16:46:18 <cdent> dstanek: did you have anything you wanted to bring up or did you show up because that's what you do? 16:48:12 <cdent> what about you elmiko? 16:48:20 <elmiko> i've said my peace 16:48:30 <cdent> If not I think we can call it, we've sort of steamed out 16:48:39 <elmiko> +1 16:48:56 <cdent> One final piece of noise: let's use email more 16:49:15 <cdent> thank elmiko, etoews and dstanek. 16:49:21 <elmiko> agreed, i'll send something about the merged guidelines 16:49:23 <dstanek> cdent: no not from me 16:49:57 <cdent> Next week is the late night non-meeting and afer that do we want a meeting in the run up to summit? 16:50:05 <dstanek> looks like i have a bit of work to do to catchup before the summit :-) 16:50:08 <elmiko> probably a good idea 16:50:13 <cdent> cool 16:50:48 <cdent> then 16:50:51 <cdent> #endmeeting