15:00:13 #startmeeting StoryBoard 15:00:13 Meeting started Mon Jul 7 15:00:13 2014 UTC and is due to finish in 60 minutes. The chair is krotscheck. Information about MeetBot at http://wiki.debian.org/MeetBot. 15:00:15 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 15:00:18 The meeting name has been set to 'storyboard' 15:00:21 Hi 15:00:21 Good morning, beautiful people! 15:00:29 o/ 15:00:38 Ok, so looking at the agenda I see two major things which I think are going to take the majority of the discussion time. 15:00:52 So I’m going to rearrange the meeting a bit and get the quick stuff out of the way. 15:01:00 (Those two items being Vision and Search) 15:01:05 o/ 15:01:12 krotscheck: link to agenda? 15:01:22 Agenda: https://wiki.openstack.org/wiki/StoryBoard 15:01:37 * krotscheck makes a note to self to make someone else update meetingbot with agenda pages. 15:01:56 #topic Ongoing Work. 15:02:13 Ish is still in HP background check hell. 15:02:22 * krotscheck is super annoyed at HR processes. 15:03:10 I spent last week with a bunch of rebases, specification docs, and getting annoyed at font encoding engines. Ultimately, little actual progress, though we did manage to land Quicknav. 15:03:34 MaxV: You around? 15:03:52 Guess not. NikitaKo_ ? 15:03:58 I'm here 15:04:18 So, I've been working on Teams API 15:04:26 * krotscheck saw that, and likes where it’s going. 15:04:27 The DB part is on review 15:04:43 The superusers migration is also there 15:04:56 and the REST part is comming 15:05:12 Excellent. 15:05:23 I've also started looking at oslo.db 15:05:47 NikitaKo_: Oh? Is that new and/or somethign we should really use? 15:05:50 and I guess we will be able to migrate soon 15:06:16 Nothing new there, but now it will be installed through requirements.txt 15:06:27 which means no code syncs 15:07:01 That’s nice. 15:07:11 that's all from my side 15:07:16 ttx: You wrote a pretty spiffy vision doc, which we’ll talk about in a bit. Anything else? 15:07:28 I iterated on Tags specs. Haven't worked on team/groups although I think the use cases I contributed to https://review.openstack.org/#/c/103667/ is probably most of it 15:07:52 so not sure a separate doc is still needed on that? 15:08:18 Fair point. NikitaKo_? Thoughts? 15:09:14 krotscheck: hello 15:09:44 krotscheck: I can give it another thought and if I have more, propose it in a comment on that spec 15:09:53 ttx: That’ll work. 15:09:54 if we think it belongs to another doc we can still wiki it 15:10:02 MaxV: Hey there! Any updates from last week? 15:10:46 krotscheck: re separate spec, I need to have a look 15:10:51 krotscheck: No, I didn't had the time due to scss introduction in Horizon, but I think I will be able to make some patches and reviews during this week 15:11:09 MaxV: Wow, sounds like y’all are doing good work over there :) 15:11:16 krotscheck: I have started a script to avoid to download node again on each run 15:11:20 Alright, Let’s see if teams and permissions can be merged. 15:11:32 MaxV: That would greatly speed up our builds. 15:12:20 MaxV: It might be intersting to look at Trusty as a build node, since the version of node on that is way more recent. I’m a few months removed from that code though... 15:12:27 MaxV: ++ 15:12:48 Alright, on to specs... 15:12:51 #topic Specs 15:13:14 So, from previous comments it sounds like Teams/Permissions are still under discussion. 15:13:41 FYI - I just added the permissions spec because it was on Etherpad and I didn’t want to lose my thoguhts there, if it becomes superfluous then we can get rid of it. 15:14:25 Most of the specs discussion (from past history) will likely focus on search, so I’d like to mention the others first. 15:15:20 First question for jeblair: At what point is a spec “approved”? For example, we’ve got the subscriptions spec, and it seems to have settled, and we’re already putting some of the pieces into place for that, but it’s not “Approved” yet. 15:15:46 jeblair: So what’s the expectation there? 15:17:27 Ok, so teams & permissions are still up for discussion. 15:17:29 krotscheck: i'll take a look at that again soon then 15:17:30 Other than that we still have search. 15:17:47 ttx, you added a comment to the search spec that I thought was interesting. 15:18:01 Quote: “The client-side implementation looks good. However game is not over yet on the server side final implementation. Maybe we could split this into initial search and scalable search, since it looks like we are implementing the first now ?” 15:18:07 krotscheck: approved is, well, approved -- so when an infra-core leaves an approval vote 15:18:14 krotscheck: and the patch lands to the repo 15:18:22 jeblair: Does it seem strange to you that storyboard-core can’t approve its own specs? 15:19:00 krotscheck: not at all. storyboard is a constituent project of the infrastructure program, and the overall direction of that program is set by the ptl and core team 15:19:26 krotscheck: i plan on seeing that you, ttx, and other important players are in agreement before approving storyboard related specs 15:19:54 Got it. 15:20:04 I disagree, but now’s not the time to talk about that. 15:20:14 On to ttx’s comment. 15:20:22 krotscheck: if i do something you don't like there, feel free to discuss it with me. you can always run for ptl or take it up with the tc. 15:20:45 I thought his comment was intersting in the split between “initial search” and “scaleable search" 15:20:52 krotscheck: and as a last resort, you're welcome to work on whatever projects you want outside of the context of the openstack project. 15:21:04 See, the client implementation is purely a browse UI. 15:21:30 There is a change for the server side 15:21:31 krotscheck: but as long as you're trying to work within it, please do attempt to follow our processes. 15:21:50 jeblair: Have I not? 15:22:19 krotscheck: you said you disagreed with them. 15:22:21 jeblair: As I said: A bit out of context for the scope of this meeting, so I’ll drop you a private email later. Honestly, I feel like most of the things will end up on the disucssion for the vision doc. 15:22:46 which can be called "initial" 15:22:50 here it is https://review.openstack.org/#/c/101476/ 15:22:59 Right! 15:23:01 Soryr 15:23:07 * krotscheck is getting split between two conversations. 15:23:16 krotscheck: i didn't start the conversation, but i am prepared to finish it. i look forward to your email. 15:23:31 krotscheck: my point there is that I expect the discussion on the final server-side solution to take a bit of time -- and it shouldn't block the client-side implementation (which uses the primitive server-side SQL-backed search) 15:23:38 jeblair: Yes please, I’ll drop you a line (I don’t disagree with process, just with underlying assumptions) 15:24:01 I totally agree with you ttx. 15:24:18 we can "improve" the server-side afterwards 15:24:24 I also agree 15:24:31 (inserts something about bridges being crossed) 15:24:47 Action item for me to split it 15:25:09 #action NikitaKo_ Split search specification into “inital” and “scaleable" 15:25:56 I like the initial abstraction on that. 15:26:23 So it sounds like initial search has consensus, do we want to talk about scaleable search? 15:26:38 * krotscheck may be wrong in assuming consensus. 15:26:56 it's fine by me (initial) 15:27:33 I don't have strong opinions on scaleable, i'll defer to those who will implement/maintain the beast 15:28:08 * krotscheck has been familiarizing himself with the guts of Lucene/ES recently, so that he knows what he’s talking about :) 15:28:23 Anyway the scalable solution will require messaging and a a lot of changes to puppet modules 15:28:50 Messaging would be driving index rebuilds, yes? 15:29:01 krotscheck: yes 15:29:23 Righto 15:29:48 I think it makes sense to use messaging for both notifications and indexing 15:30:05 NikitaKo_: I agree. 15:30:44 I’m a little worried that my own self-investment in that spec (due to my authorship) is making me blind to other potential solutions though, so if anyone has different thoughts... 15:31:00 * krotscheck doesn’t want messaging to be the hammer and everything else to be a nail. 15:31:42 Anything else on specs before we switch over to Vision? 15:32:23 #topic Vision Document (ttx) 15:32:25 https://wiki.openstack.org/wiki/StoryBoard/Vision 15:32:28 OK, so I drafted a basic vision while you were celebrating independence, trying to stay high-level and hopefully implementation-agnostic 15:32:47 I wrote remarks at the bottom 15:33:01 about things that were not as self-evident as the rest 15:33:23 which may trigger more discussions imho 15:33:34 As it should :) 15:33:46 Not sure if the rest of the "vision" is fully consensual though 15:34:00 it's just that it flowed consistently when I wrote it 15:34:20 while the two sticky points felt more artificially attached 15:34:47 I still stand by those, because I still think it's the easiest way to solve that hole in our tooling 15:34:54 but I welcome discussion on them 15:35:35 Honestly, I love that doc because I can pull actual features out of it very quickly. 15:35:37 it's a first draft, comments welcome 15:35:56 Case and point: The first paragraph tells me that tasks do not necessarily have to be attached to code. 15:36:04 krotscheck: did I manage to stay implementation-agnostic enough ? 15:36:18 The only thing that's pretty hardcoded in there is the task/story relationship 15:36:30 ttx: Yes. I see no python. 15:36:32 :) 15:36:58 The third paragraph also tells me that personal goals are a separate way of organizing tasks. 15:37:22 I have one question: Does the spec meaningfully change if we replace “gerrit” with “code review system”? 15:37:43 (similar with git vs. source control)( 15:38:01 krotscheck: no. we could say code review system and VCS 15:38:06 +1 15:38:11 * ttx fixes 15:38:59 Which of the notes do you want to talk about? 15:39:06 The bug vs task tracker is interesting. 15:39:22 We have the “is_bug” flag in the db righ tnow. 15:39:55 so, another thing that emerged from this is that we could have 3 types 15:40:02 feature, bug and vulnerability 15:40:14 which are the key components of changelogs/release notes 15:40:26 Interesting distinction 15:40:27 taking a step back allowed me to see the 3rd type 15:40:58 since I suddenly no longer saw it as a bug, but as a piece of release information to convey 15:41:22 Hrm. 15:41:25 that said, if we drop the changelog task and let the VCS build it for us, it's a non-issue 15:41:28 Iiiiiinteresting. 15:41:44 back to bugtracker/tasktracker 15:42:09 the problem with bug tracking is that it's almost a workflow that happens before the main use case (task tracking) happens 15:42:26 we can make the task tracker bug-submission friendly (which is the way we started to do that) 15:42:46 or we can try to have a separate tool / separate panels for bug submission/triaging 15:43:12 I still think the first option is the simplest, but i'm ready to have a long thought about it 15:43:39 I know that particular bit was bothering gothicmindfood as well 15:43:49 That kindof jives with some of the UX feedback I got around the workflow for a feature. Design discussion is something that happens well before any tasks are assigned, right? 15:44:28 krotscheck: that's one way of seeing it yes... You could also consider it's just one of the tasks 15:44:33 True 15:44:37 task 0 15:44:47 especially now that it lands in a code repo 15:44:48 * krotscheck wonders if a bug is simply a story with the Bug system tag 15:44:51 and we can track the status of that 15:45:02 Hi there, ish_! 15:45:18 Hi Michael.. 15:45:43 krotscheck: the trick there is... developers create stories and tasks for themselves in most cases. They may look at bugs and triage them, but it's almost a separate workflow 15:46:08 ttx: Workflow-by-tag? 15:46:19 * krotscheck is just throwing concepts up against the wall here. 15:46:21 krotscheck: yeah, probably the simplest 15:46:38 krotscheck: alternative is to have an incident response component in parallel of the task tracking component 15:46:53 but I fear that the former would get ignored 15:47:08 ttx: Not if we surface it. 15:47:12 given that our devs are all cats 15:47:14 * gothicmindfood might be tired, but often thinks bugs are just a tag 15:47:16 I like cats 15:47:31 I have two of them. As long as you have a little shiny red dot they will go wherever you want them to. 15:47:47 krotscheck: anyway, I just wanted to single out that those two parts of the "vision" were probably the ones were discussion should still happen 15:48:10 * gothicmindfood also requests a joke 'feature' tag to put with bugs if someone wants 15:48:10 do you see anything questionable in the rest of the vision ? 15:48:26 Nope. 15:49:13 doesn't mean it's done or designed -- the whole "organize their work" thing is pretty nebulous at this point 15:50:18 ttx: Well, given that we can (soon) describe a search via a fixed set of criteria, saving those criteria into your own customized dashboard isn’t hard :) 15:51:44 ttx: Ok, so what part in particular do you want more focus on? 15:52:12 krotscheck: do you mean we need something like gmail has, when you do a search and then say "create me a filter by that search" ? 15:53:19 NikitaKo_: That’s what I’m thinking. Rather than try to proscribe what goes on someone’s dashboard, let people design it themselves. 15:53:33 (and then go in and figure out what people have actually built, and inform our use cases from that) 15:54:00 But that’s pie-in-the-sky right now. 15:54:16 #topic Open Discussion 15:54:17 krotscheck: I think the focus should still be on the milestones we defined on the roadmap 15:54:23 users could build their subscriptions with the same mechanizm then 15:54:23 ttx: I agree 15:54:48 krotscheck: defining them after user adoption requirements is a pretty fgood way of doing it 15:55:07 NikitaKo_: Yup. 15:55:27 Ok, so we have 5 minutes, are there any major roadmap blockers that we’re running into? 15:56:03 One thing I noticed is that we have lots of admin-related tasks but we don’t have any place to really put them into the UI. 15:56:29 Oh, also, ttx: Are project groups an admin thing or a anyone thing right now? 15:57:06 krotscheck: I think it can be an admin thing 15:57:17 coupled with personal project subscription 15:57:33 should cover most cases... 15:57:57 ttx: Admin it is then. 15:58:07 Nobody has any roadmap blockers/ 15:58:08 ? 15:58:14 (Other than code reviews? :) ) 15:58:45 nothing that I can see 15:59:00 nop 15:59:36 Awesome. 15:59:40 #endmeeting