21:02:56 #startmeeting crossproject 21:02:56 yo 21:02:58 Meeting started Tue Jan 12 21:02:56 2016 UTC and is due to finish in 60 minutes. The chair is thingee. Information about MeetBot at http://wiki.debian.org/MeetBot. 21:02:59 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 21:03:01 The meeting name has been set to 'crossproject' 21:03:02 o/ 21:03:06 o/ 21:03:07 * rockyg is still sleepily peeking out from under that stone.... 21:03:17 sorry for jumping people around. Just making sure they know about this new irc channel for the cross project meeting 21:03:22 heh 21:03:30 no worries thanks thingee 21:03:35 indeed, i almost forgot after several weeks of no agenda 21:03:52 we have a lovely agenda 21:03:55 https://wiki.openstack.org/wiki/Meetings/CrossProjectMeeting 21:04:11 since we have no past items 21:04:24 #topic team announcements (horizontal, vertical, diagonal) 21:04:43 o/ 21:04:45 openstack security project meetup is this week 21:04:52 and barbican 21:05:01 do we have a wiki page with all of the meetups? 21:05:07 yep, sprints 21:05:19 #link https://wiki.openstack.org/wiki/Sprints 21:05:28 thank you 21:06:21 * jroll lurks 21:06:23 just a reminder of dhellmann's release countdown http://lists.openstack.org/pipermail/openstack-dev/2016-January/083507.html 21:06:45 mitaka 2 coming jan 19-21 21:06:54 time flies 21:07:08 seriously... 21:07:09 #link http://docs.openstack.org/releases/schedules/mitaka.html 21:07:10 when you're having fun 21:07:19 ok going, going 21:07:25 oh are we here now 21:07:44 welcome ttx! 21:07:57 #topic API guides vision - developer.openstack.org and REST API docs 21:08:02 hey 21:08:04 annegentle: hi 21:08:25 I wanted to make sure I set up the current situation with application dev docs, so that project teams know what's up. 21:08:44 in keystone we have the v3 spec is just in rst 21:08:57 it's an exciting time with 900 API calls in OpenStack that are documented. We now have over 30 REST API services, but 12 projects that have API reference info 21:09:10 #link http://developer.openstack.org/ 21:09:26 We have a new landing page design, and we're starting to get even more how-to content which is super exciting. 21:09:45 bknudson_: yeah that's part of this launch also, is getting out of Docbook/WADL to RST/Swagger. 21:09:53 * stevemar sneaks into the back of the room to listen 21:10:04 annegentle: sahara is working towards this with out proposed v2 api 21:10:09 s/out/our 21:10:11 So, we need more how-to, we need centralized locations for both the content and the reviewers who know what's up for this audience, and we need standards. 21:10:22 Ask and ye shall recieve. 21:10:25 receive? 21:10:27 I can't spell. 21:10:31 Anyway. 21:10:45 #link http://www.openstack.org/blog/2016/01/whats-next-for-application-developer-guides/ 21:11:14 do you have an automatic conversion from wads to swagger? 21:11:18 wadl 21:11:27 that post outlines a lot of the vision -- and maybe you got a chance to read it before coming here -- I wanted to be available for any questions 21:11:39 bknudson_: just last week we brought in fairy-slipper, a migration tool 21:12:06 #link https://review.openstack.org/#/q/project:openstack/fairy-slipper 21:12:32 annegentle: so, the docs team doesn't want to add projects to the install guides and such right now... just to be sure, this is open to all projects? 21:12:40 This tool had a proof-of-concept last release, and now we're collaborating more widely with the API working group. 21:12:45 really what we need to do is convert our v3 api spec to swagger -- http://specs.openstack.org/openstack/keystone-specs/api/v3/identity-api-v3.html 21:12:53 annegentle: i'm curious about adding links to the api-wg wiki and guidelines on developer.os.o, wondering if adding another section to the API portion would be appropriate 21:13:17 jroll: yes, there isn't as heavy a testing requirement on REST API and app dev docs 21:13:27 annegentle: cool, ty :D 21:13:27 we've got "References" now, but about a "Working Group" section or something? 21:13:28 elmiko: sure 21:13:57 elmiko: it'd be great to rework that landing, but the first step was to get the landing page looking closer to a real web design :) 21:14:07 right 21:14:26 We won't do all the WADL migration at once, so teams should continue to update WADL in api-site so that we know to migrate it 21:14:46 teams can also investigate building how-to documents for API guides similar to what Compute is doing 21:14:53 will the individual teams eventually be responsible for creating their own swagger? 21:14:59 #link http://developer.openstack.org/api-guide/compute/ 21:15:10 elmiko: yes, that's idea, since the teams can review the design the best 21:15:16 elmiko: while following api-wg guidelines 21:15:21 annegentle: awesome, thanks 21:15:29 source for that api-guide is in the nova repo 21:16:22 oh, the source goes in the project tree, awesome 21:16:23 Teams should generally be aware of developer docs helping out their application devs immensely. 21:16:33 jroll: cool, glad you see the vision :) 21:16:56 We 21:16:56 Maybe it would be worth it to start fresh with the identity docs rather than convert the wadls, since keystone team hasn't been looking at wadls. 21:17:22 bknudson_: yea, i almost feel the same about sahara. i'm not sure how far behind the wadls are 21:17:39 it's interesting, we had over 120 contributors to the WADLs last release 21:17:47 nice 21:17:51 so even if you think they're not updated, they're the most contributed to docs as of late 21:18:31 our most recent docs core contributors worked mostly on the API reference docs, KATO Tomoyuki. 21:19:04 we don't want to have 2 sources of truth for the api spec. 21:19:12 so I think migration's the way to go, unless a team really has a great Swagger file. For example, me and a coworker wrote one for Magnum since they didn't have a WADL anyway. 21:19:18 bknudson_: for sure 21:19:54 now, to jroll's point, we are starting with the six infra-services for migration purposes 21:20:21 Identity, Compute, Images, Networks, Block Storage, Object Storage 21:20:33 since we need to test the migration 21:20:44 Additional needs and calls to help: 21:21:26 We need to know the WADL-to-Swagger captures the API definitions complete as possible. That's fairy-slipper work. 21:21:43 We need to publish Swagger plus RST using flat files and file copy similar to how docs jobs work today. 21:22:03 We need build jobs and content in project's repos for how-to and tutorial examples for how to use your REST APIs 21:22:16 That's all I've got! 21:22:31 the how-tos are in the project repo and the swagger is elsewhere? 21:22:34 Feel free to ask me anything in #openstack-sdks or #openstack-doc and please check out fairy-slipper 21:23:10 bknudson_: yes, we have a couple of ideas for further integration, and ideas for generating swagger, but have to take small steps that give continuous deliverables to this audience. 21:23:16 annegentle: so where do we go from here to get people involved from their respected projects? 21:23:39 thingee: each project should have an API liaison and I can reach out directly to them 21:23:52 excellent 21:24:14 anything else for annegentle before we move on...? 21:24:16 Also, people need to write tutorials! There are some cool things you can make with OpenStack infrastructure :) 21:24:21 does the doc team meeting usually cover the topics related to swagger and descriptive api doc gens? 21:24:31 annegentle: should we add a section to the CPL list for API Liaison work? 21:24:40 elmiko: yep, we have a standing item on the agenda as a "specialty team" 21:24:48 annegentle: awesome, thanks 21:25:02 nikhil: that would be great, I was just going to look for that. A while back we had API specialists, but I'm not sure that list is maintained. 21:25:16 gotcha 21:25:17 what about the api-wg liaisons? 21:25:18 Pretty sure it's not. :) 21:25:26 I am thinking this page for tracking 21:25:29 #link https://wiki.openstack.org/wiki/CrossProjectLiaisons 21:25:35 I would be amiss if I didn't thank people like Russell Sim and Karen Bradshow for getting us this far! 21:25:52 elmiko: that's the list I was thinking of, api-wg liaisons! 21:25:59 \o/ 21:26:03 elmiko: do you have that list? 21:26:03 thanks annegentle! 21:26:19 #link http://specs.openstack.org/openstack/api-wg/liaisons.html 21:26:26 #info Many thanks to Russell Sim and Karen Bradshow for getting us this far! 21:26:43 elmiko: fancy json no less! Nice. 21:26:54 we used to have a nice table in the wiki, not sure what happened to it 21:27:10 #topic Cross-Project Spec Liaisons 21:27:36 by default these should be ptls 21:27:53 ok so so some time ago I spoke on the idea of having representatives from each project who can help with cross-project specs 21:27:55 http://eavesdrop.openstack.org/meetings/crossproject/2015/crossproject.2015-12-01-21.00.html 21:27:57 #link http://eavesdrop.openstack.org/meetings/crossproject/2015/crossproject.2015-12-01-21.00.html 21:28:02 fungi: definitely 21:28:07 at least that's how our other liaisoning is done (i think i just made up a word) 21:28:40 liasing I think 21:28:44 unfortunately expressed at the summit ptls don't always have time for this meeting or the efforts happening in cross-project areas. Alarming I agree, but I need someone from teams to participate and get ptls involved when necessary 21:28:54 fungi liasing 21:28:56 is the word ;) 21:28:58 lifeless: +1 21:29:15 anyway, consider infra to be ptl-liaisonated for now, and i'll gladly delegate i it becomes necessary 21:29:28 that leads me to the first attempt of defining this group 21:29:30 #link https://review.openstack.org/#/c/266072/3 21:29:31 liaisonated is definitely a word ;) 21:29:37 it should be =) 21:29:37 as stated ptls are default. 21:30:04 it will be listed in https://wiki.openstack.org/wiki/CrossProjectLiaisons for someone else to sign up 21:30:11 thingee, maybe a weekly crossproject posting of open reviews, summary/title and status? Then lots of people would get a gentle reminder every week 21:30:18 but the responsibilites are more or less of what we discussed in that previous meeting. 21:30:21 I thikn I managed to volunteer or get volunteered for nova 21:31:25 assuming we can agree on the responsibilities in the project team guide, I will announce on the ML the signup table for each team 21:31:29 cdent: thank you! 21:31:46 (as the guy who comes here) 21:31:58 so with the 'big-tent' there are "who knows how many" liasons/ptls. i'm not sure how successful 'all you people look at this' approach is. 21:32:03 thingee: sounds great 21:32:04 i'd like to talk with our ptl (sahara), but i'm certainly up for liasing on sahara's behalf 21:32:55 elmiko: wonderful 21:33:09 gordc : it doesn't have to be everyone, but it needs to be more people than are reviewing them now 21:33:11 i would think being cross project, the scope of specs requires quite a detail knowledge of not just one's own project but others 21:33:20 thingee: what tag should we expect on the subject? 21:33:20 thingee: we could mention the open specs in the weekly dev digest too, I guess 21:33:28 yeah ideally I will mention it on the list and then start bugging individual project meetings to announce it who aren't having someone from their team attend this meeting 21:33:31 gordc: i think it reflects on those projects well or poorly depending on how involved they get in the cross-project specs discussions and implementations 21:33:33 +1 ttx 21:33:39 ttx: I did that once or twice 21:33:40 fungi : ++ 21:33:46 ttx: I will continue though :) 21:34:01 fungi: well there's also resource constaints 21:34:12 not all projects have 100s of devs like nova. 21:34:15 nikhil: cross-project 21:34:15 bandwidth... 21:34:21 gordc, if you or the project liaison covers your project's view and every project comments, then you don't have to worry about knowing other projects 21:34:22 thx 21:35:06 gordc: the work for a smaller project is also likely to be more reasonable, but I see your point 21:35:08 gordc: yeah so I have a bullet that addresses that... I think for implementing, that's a different story. But it needs to be prioritized by the liaison bring it to the groups attention and ptl 21:35:11 gordc: of course, resource-constrained projects already have plenty of things to cause them strife. i expect projects who can't keep up with their own work are likely to struggle with following cross-project specs too 21:35:22 rockyg: i guess my assumption is that there will be specs that cannot be silo'd to just a single project view. maybe a future concern 21:35:48 fungi: ++ 21:36:29 ok well again please comment on the review with suggestions. https://review.openstack.org/#/c/266072/3 21:36:39 fungi: does anyone truely keep up with their own work ? 21:36:56 also, if all (most) projects comment and there are conflicting views, that's the start of an important discussion for all the involved projects 21:36:56 I'd like to get this going start next week with getting initial liaisons so we can help those work on specs to bring attention 21:36:57 thingee: i'll comment there. i'm just being negative :) 21:37:00 * lifeless goes off the deep philosophical end 21:37:05 lifeless, ++ 21:37:07 lifeless: probably not, some just get closer than others 21:37:08 lifeless: lol, so true... 21:37:14 and so projects are aware of what's going on from a cross-project perspective 21:37:32 rockyg: yeah, that's another concern. how to get agreement with so many possibel voices... 21:37:38 anything else before we move onto the next topic? 21:37:47 nothing from me 21:38:07 i guess my point was that projects will prioritize cross-project spec work differently, so it's not necessary that you prioritize it above, say, getting your project working in the first place 21:38:23 thingee: ttx: have you considered including the product working group or other working groups or asking them to review these? 21:38:46 annegentle: absolutely. I work with the product working group, so this sort of plays into their goals as well 21:38:52 annegentle: anyone can review these, and I think thingee has started more actively reaching out 21:39:01 thingee: ttx: ok cool 21:39:11 esp thinking of the themes and getting Carol involved 21:39:28 really, I feel like we need to spread the workload 21:39:30 carol is going to be out for a bit 21:39:34 and perspective 21:39:57 #topic open discussion 21:39:58 like yummy peanut butter 21:40:04 announcement 21:40:21 The OpenStack Mentoring program sponsored by the women of openstack is looking for mentors! 21:40:41 oslo has decided to adopt the os-profiler project 21:40:54 we need people with experience and we're looking at doing a workshop a day before the summit 21:41:04 thingee: following the tc meeting it appears we'll have to document how the tc members are supposed to interact with cross project specs 21:41:09 experience with contributing through translations, docs, code 21:41:20 thingee: Is there an etherpad or something with more details? 21:41:25 signup https://docs.google.com/forms/d/1YQITea0ygvusZAeRdR6jL1VUfaNcZTCr2UqHTFSsxR4/viewform?c=0&w=1 21:41:40 thingee: Thanks! 21:41:47 dims, ++ 21:42:09 Hopefully we can dedicate sometime before the summit to helping our community grow healthy 21:42:38 +1 21:42:42 dims: oh interesting, I think I missed that on the ML 21:42:56 exercises for people huddled on the floor with their laptops ??? 21:43:13 * elmiko chuckles 21:43:13 hopefully I'm not dense, but looking for mentors of all genders right (and it's just *sponsored* by Women of OpenStack) right? 21:43:26 if people want more information on the OpenStack mentoring program please contact ekhugen on #openstack-women 21:43:38 docaedo: oh yes, mentors can be anyone 21:43:40 docaedo: good question, yes all genders are welcome! 21:43:53 annegentle: what's the scope of this program? 21:44:05 is this different from official internship? 21:44:11 please spread the word too if you can't participate. Every bit helps 21:44:13 nikhil: this isn't Outreachy, this is new, right thingee? 21:44:35 nikhil: career mentoring or technical mentoring 21:44:45 gotcha 21:44:45 thingee : spec merged just today, will drop a note on ML 21:45:11 correct this is different from outreachy. It has no set schedule like outreachy, and it's a way to have a pool of mentors accessible to people who need help contributing 21:45:46 aside from the scheduled in person mentoring at the summit ;) 21:45:49 so not funded internships, more a mechanism for connecting people who want to learn with people who want to help them? 21:45:58 fungi: yes 21:46:08 fungi: thank you 21:46:25 anything else? 21:46:34 excellent 21:46:58 ok, comment on https://review.openstack.org/#/c/266072/3 now and thanks everyone! 21:47:04 I have one 21:47:05 yay cross-project \o/ 21:47:10 * thingee stops 21:47:15 there's a very likely "openstack hackathon" global event happening in the week of feb 29 21:47:16 nikhil: yes 21:47:33 ooh, neat 21:47:44 nikhil: cool! 21:47:49 * thingee every week is a hackathon 21:47:51 currently planned in east & west coast US, one location in UK and one in China 21:47:57 Yeah. It's an outgrowth of the China Hackathon. It's for bug fixes 21:48:11 I am working on one location in India 21:48:22 nikhil, ++ 21:48:25 link details? 21:48:35 I need help getting estimate of interest from people 21:48:55 We are coming up with it as logistical bottlenecks are being solved 21:48:56 It's likely Huawei will host in SF Bay area 21:49:06 Seemed like a good place to declare 21:49:11 rockyg: nice 21:49:23 there's a texas location too 21:49:25 afaik 21:49:32 So, we'll need lots of cores ready to review the fixes...and others 21:50:00 yeah get the dates and locations published as soon as you can 21:50:08 Maybe PTLs and cores could work up the list of bugs they want folks to focus on? 21:50:14 hopefully the participants will also do reviews 21:50:37 nikhil: I think it would be good if there it's a signup form/etherpad and mention on the ML 21:50:45 I'm just now hearing about this anyways :) 21:50:53 thingee: surely, I will try to get one before next mtg 21:50:55 Definitely. Part of this is so folks with different languages can do most of the discussion in their own languages. 21:51:17 ok thanks everyone! 21:51:22 I would like to work with the mentoring group to get people involved in this 21:51:23 #endmeeting