22:01:24 #startmeeting horizon 22:01:25 Meeting started Tue Oct 1 22:01:24 2013 UTC and is due to finish in 60 minutes. The chair is gabrielhurley. Information about MeetBot at http://wiki.debian.org/MeetBot. 22:01:27 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 22:01:29 The meeting name has been set to 'horizon' 22:01:32 hello folks 22:01:36 _o/ 22:01:37 hello 22:01:37 Hello 22:01:39 hi 22:01:42 o/ 22:01:45 #topic overview 22:01:53 https://launchpad.net/horizon/+milestone/havana-rc1 22:01:53 hi all! 22:01:58 that's pretty much the story, there 22:02:06 last five bugs, all with patches that are merging presently 22:02:23 keep an eye on 'em to make sure they don't get stuck or anything, but that's where we're at 22:02:40 hello 22:02:54 once they're all merged I'll push a version bump, ttx will cut the RC from that commit, and master is open for icehouse work 22:03:16 after the RC is cut, any fixes that need to go into Havana need to be fixed in master first, then backported to the RC branch 22:03:33 only if an RC window is reopened 22:03:38 (the backport) 22:03:41 indeed 22:03:50 best case is we don't have any need for an RC2 22:03:55 i.e. we need to have an annoying bug at least 22:04:18 given how much work we've done to vet RC1 I think it's possible we won't, but better to know what happens if we do have to do an RC2 than not. 22:04:56 generally the bar for respinning the RC for me is either a security problem or a significantly broken user-facing feature. 22:05:44 The one last thing I will do today is to write up the release notes and commit those to the repo, as is my custom. that'll come in right before the version bump commit 22:06:26 Aside from the RC, the only other official business this week is that PTL elections are currently in progress. You should have received an email with a link to vote. If you did not, please let me know. 22:08:06 I don't have any other topics for this meeting. The bugs are in a known place, and it's a little early to start reviewing blueprints for Icehouse. I'm just gonna skip straight to open discussion. 22:08:09 #topic open discussion 22:08:17 free for all! 22:08:45 The bug mentioned in the last meeting https://bugs.launchpad.net/horizon/+bug/1211535 is not really complete 22:08:47 Launchpad bug 1211535 in horizon "HorizonMiddleware class causes excessive session creation" [Medium,Fix committed] 22:09:12 gabrielhurley: question, do you got a clue on the resource introspection stuff and realtime notifications ? 22:09:14 the fix that went in django_openstack_auth was a minor fix that complemented it 22:09:43 david-lyle: no? 22:09:52 yeah, I thought the other half was also merged 22:10:13 the patch to Horizon was reverted because the fix in django_openstack_auth was needed, the original patch has not been reposted 22:10:20 ekarlso: introspection is an Icehouse maybe. realtime is an icehouse definite. 22:10:45 gabrielhurley: isn't introspection already due date ? 22:10:55 gabrielhurley: we can close Havana without it, but we should reopen it 22:11:03 i mean, i thought you talked about it for grizzly ? 22:11:05 david-lyle: ah, I missed that it hadn't been re-posted. let's get that up now that django_openstack_auth is published. 22:11:58 ekarlso: that's true, but there's no current resources available to tackle it. it was one of my personal projects, but my budget for coding time on Horizon ain't what it used to be. the realtime stuff has a lot more interest and energy around it right now, and it's one I'm currently more invested in. 22:12:16 It's only about 50% true that what I say "just happens" 22:12:18 ;-) 22:12:35 gabrielhurley: lots of nebula stuff going on rather then horizon ? ;p 22:12:51 david-lyle: does the patch need any changes or can it just be re-approved as it was with the new openstack_auth? 22:12:53 gabrielhurley: if we do, we'll have to bump openstack/requirements to 1.1.3 for django_openstack_auth and then horizon as well. Causes ugly error without django_openstack_auth fix 22:13:28 hmmm... the upper range isn't pinned, so it'd get picked up in new ones, but not in old ones. 22:13:59 but, Horizon can't pin unless openstack/requirements pins 22:14:16 okay, I re-opened the ticket. I'll tag it as a havana-rc-potential and we can revisit it *if* we need to cut another RC. let's get the work done on it though. 22:14:33 ok, yeah, I think it's an all or none kind of deal in my mind 22:14:37 if by chance it all lands before everything else in the gate, that simplifies things, I suppose 22:15:05 gabrielhurley: sounds good 22:15:19 okay. ticket marked as such. 22:15:50 other topics? 22:16:15 i think https://review.openstack.org/#/c/42240/ was the repost fwiw 22:17:17 I'll work with the author to iron this out 22:17:26 thanks 22:19:43 going once.... 22:19:46 I've created some wireframes around some details pages for Tuskar... 22:19:49 http://people.redhat.com/~lsurette/OpenStack/Tuskar%20Detail%20Pages_1.0.pdf 22:19:55 if anyone has interest in reviewing them 22:20:03 there is a thread on the ML 22:20:12 cool 22:20:36 lblanchard: at first glance, those look really nice, but no tables !! :) 22:20:36 even though these are focused on Tuskar, I think there could be a lot of reuse in the visualizations for Horizon proper 22:20:50 david-lyle: true, no tables :) 22:21:09 david-lyle: but a user would probably get to these from tables, haha 22:21:18 gotta have tables 22:21:25 david-lyle: maybe there is information missing that belongs in tables 22:22:27 those are some nice visualizations lblanchard 22:22:41 very nice. my one piece of practical advice on these (from having gone through this with Nebula's designer working on both OpenStack and Nebula's dashboard) is that the "ideal"/"end goal" wireframes are often so far out that even beginning to implement them can be tough. Data sources for many of those metrics are difficult and deployment-dependent. Beyond that, I can identify a dozen things in those that don't exist in 22:22:55 I guess what I'm trying to get at is that it might also be helpful to have "transition" wireframes 22:23:17 timductive1: thanks! 22:23:19 or some plan for how we achieve this end goal, and all the stakeholders involved 22:23:21 gabrielhurley: fair point 22:23:36 right.. I have actually some wireframes around tuskar as well 22:23:37 also check that d3 can handle all of those visualization types 22:23:39 gabrielhurley: I agree…I'll be working closely with lsmola on this 22:23:46 mostly around processes, which might help to get the idea better 22:23:51 let me search for the links :) 22:23:52 jcoufal: I was gonna ask how much you and lblanchard's work overlapped here... 22:24:22 timductive1: it can with a little elbow-grease. You can build whatever you need with it. 22:24:27 gabrielhurley: jcoufal and I have been working closely…making sure we don't step on each others toes, and being consistent with designs 22:24:28 gabrielhurley: lblanchard is focusing more on detail pages and metrics, I am more dealing around concepts and workflows 22:24:40 I've seen many of those things built before, which is why I know the cost in work-hours. 22:24:42 we are in sync, reviewing and cooperating 22:24:57 lblanchard, jcoufal: great. just making sure. 22:25:19 gabrielhurley: :) 22:25:40 gabrielhurley: I will be sure to help development through transitional wireframes if information isn't available to start... 22:25:49 here are latest proposals 22:26:01 #link http://people.redhat.com/~jcoufal/openstack/tuskar/2013-09-30_tuskar_l-group_creation_wireframes.pdf 22:26:05 in general I totally love the wireframes. I just want to make sure we don't kill ourselves trying to get there. :-) 22:26:06 timductive1, there is also rickshaw library for linecharts 22:26:16 #link http://people.redhat.com/~jcoufal/openstack/tuskar/2013-09-30_tuskar_resource_class_creation_wireframes.pdf 22:26:24 I'm hoping Tuskar is responsible for most of the data aggregation 22:26:28 gabrielhurley, we will start with little pieces :-) 22:26:56 and the whole youtube stream explaining them :) (http://www.youtube.com/watch?v=m3y6uD8yKVQ) 22:27:03 Ismola true 22:27:33 It's also worth noting that, while Tuskar/Triple-O *is* cool and offers awesome possibilities, it's still not gonna even graduate for another 6 months at best and would be in the integrated release two cycles from now. That's a pretty long time. 22:27:36 gabrielhurley: yeah, agreed. Now that I've put together what we would love to have, I will step back and with lsmola figure out what is feasible for a first cut. 22:27:49 gabrielhurley: yes 22:28:14 gabrielhurley, we can use Horizon in tripleo now 22:28:33 gabrielhurley, so we would like to make all chart libraries upstream first 22:29:00 david-lyle: I believe the plan is that tuskar would do the aggregations…lsmola? 22:29:03 for sure. I'm not here to direct where the energy of the UX group goes. Only making sure they have full information for where prioritizing their time. 22:29:15 ignore my bad grammar 22:29:15 Another update is on navigation enhancements - will be finishing prototype, sending it out shortly and I hope we send some initial patches soon as well 22:29:34 jcoufal: I look forward to that 22:30:05 david-lyle, lblanchard yeah in Undercloud Tuskar will make the aggregations 22:30:11 side note…I'd be happy to hear from anyone if they do have certain UX things they'd like to see happen :) 22:30:45 david-lyle, lblanchard in over cloud it will be project detail pages, instance detail pages, maybe host aggregates .. 22:31:19 * david-lyle feeling better 22:31:55 my biggest UX wish is that we could roll the network topology stuff and the instance flows into one so that the topology view was the main way you interacted with the system... there's a lot of code peril to work through there, but I think it'd make the experience much more compelling. 22:32:04 david-lyle, overcloud and undercloud have similar aggregates :-) 22:32:27 kinda like the heat topology view + network topology + instances + volumes all in one go. 22:32:37 gabrielhurley: yeah, I am already got in touch with Toshi around network topology 22:32:49 I'm hoping to get time to work on a similar view for identity 22:32:51 sorry for grammar 22:32:56 gabrielhurley: great to know! Maybe I can help wireframe there too. 22:33:02 the code side is one thing, but I think having some clear wireframes demenstrating all the flows before we even start on the code would be a good idea. 22:33:10 make groups/roles/projects/domains understandable 22:33:13 it's also a bit tricky 'cuz you have to account for with *and* without Neutron 22:33:23 david-lyle: ++ 22:33:33 david-lyle: ++ 22:33:57 david-lyle: that's great, we should jump into that as well 22:34:23 between these topology type views and a real workflow, I think we can make Horizon a lot more accessible for a wider range of users 22:34:41 * david-lyle looking forward to Icehouse 22:34:41 sounds great to me!! 22:34:51 great 22:34:55 the group/role/project/domain discussion may fit in our IA session at the summit, but we may want to have a little informal hallway session on that topic alone. 22:35:11 sounds like a good idea 22:35:23 gabrielhurley: me and lblanchard want to send out initial proposal before that 22:35:28 to kick some discussion around 22:35:42 and get people to start thinking on that 22:35:42 I've tried various alternate ways of grouping that information and have had mixed success. I seem to always have to gloss over one variable to make it understandable. 22:37:31 it's difficult and I think limiting the amount of information initially visible will help, because you can make it a real nightmare 22:37:43 yep 22:38:08 still working through the layout conceptually which is a big reason I haven't started the concrete work yet 22:38:29 I've found the key to be to separate the problem spaces between the day-to-day "I need to operate in this context" and the very infrequent "I need to change something about my contexts" 22:38:48 yeah, I think writing out the use cases would be really helpful 22:38:58 and which persona would be performing them... 22:39:01 finding the right UX metaphor for "operating in a context" and "changing contexts" is a fun challenge 22:39:16 haha 22:39:17 I'm happy to help with the use cases 22:39:20 s/changing/switching 22:39:32 just need better to understand what problem we're trying to solve... 22:39:55 thanks julim 22:40:11 np david-lyle 22:40:13 it's also interesting to think about how you would do this on a touch device, because those spacial/gestural metaphors are useful in thinking about this problem 22:41:30 gabrielhurley: yeah, we were just talking to other day about mobile support…I responded to a blueprint a few weeks back but haven't seen much action on the Horizon mobile BP 22:42:03 gabrielhurley: but in general it would be great to know what a user would do from a mobile device…research needed! 22:42:14 did the assignee/proposer make any contact after the summit? 22:43:02 david-lyle: once, early on, but that was about it 22:43:08 they vanished pretty quick 22:43:26 ahh, that's what I feared, large task 22:43:59 gabrielhurley: touch devices are definitely interesting challenge, on the other hand, I think that more important would be to increase UX of dashboard in general, we can keep touch devices in mind 22:44:00 yep. I'm actually trying to recruit my good friend at Piston to do some work in making Horizon's stylesheets responsive 22:44:35 but we won't hit that much audience with touch devices as we can touch with classic UI improvements 22:44:51 +1 22:45:01 jcoufal, lblanchard: my intention there was not to suggest that we should be building *for* touch devices, but to suggest that spatial metaphors are good for representing context switches, and touch devices are an easy touch-point (no pun intended) for that mindset. 22:45:34 gabrielhurley: understood. Sorry it got me off on a tangent :) 22:45:44 no worries 22:45:55 mobile is also a great UX challenge 22:45:56 gabrielhurley: right, that's true 22:46:20 sorry for missunderstanding 22:46:23 all good 22:49:15 okay. anything else? 22:51:25 cool! I'm gonna go write release notes. 22:51:35 When next we talk we'll be looking at Icehouse 22:51:38 have a great week! 22:51:40 cool 22:51:41 #endmeeting