17:02:40 <zehicle_at_dell> #startmeeting Refstack
17:02:41 <openstack> Meeting started Thu May  1 17:02:40 2014 UTC and is due to finish in 60 minutes.  The chair is zehicle_at_dell. Information about MeetBot at http://wiki.debian.org/MeetBot.
17:02:42 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
17:02:45 <openstack> The meeting name has been set to 'refstack'
17:02:55 <zehicle_at_dell> roll call!
17:02:57 <zehicle_at_dell> o/
17:02:59 <rockyg> thanks for starting.
17:03:07 <rockyg> o/
17:03:51 <zehicle_at_dell> davidlenwell, you here?
17:04:21 <catherineD> Here
17:05:25 <zehicle_at_dell> I don't see David online
17:06:09 <rockyg> yeah.  looks like just a bot.
17:06:23 <zehicle_at_dell> let me look at the backlog
17:07:21 <zehicle_at_dell> looks like we have some docs and specs that need reivew (still)
17:07:55 <zehicle_at_dell> I can give a tcup update
17:08:09 <zehicle_at_dell> #topic TCUP
17:08:21 <zehicle_at_dell> #topic Agenda
17:08:28 <zehicle_at_dell> let's get an agenda together first
17:08:30 <zehicle_at_dell> TCUP
17:08:38 <zehicle_at_dell> DefCore status/update
17:08:42 <zehicle_at_dell> Summit Prep
17:09:07 <rockyg> Refstack dev focus short term
17:09:19 <zehicle_at_dell> Controlling critical JSON asssets (must-pass test lists & driver test json)
17:09:51 <rockyg> summit meetups (QA, others?)
17:11:32 <zehicle_at_dell> ok, let's get started then
17:11:33 <praveen_dell> Hello all
17:11:45 <zehicle_at_dell> #topic Refstack dev focus short term
17:12:32 <zehicle_at_dell> rockyg, your topic...
17:12:52 <rockyg> yo
17:12:58 <zehicle_at_dell> as far as I know, it's getting the upload API spec nailed and then posted
17:13:10 <rockyg> ok.
17:13:50 <rockyg> I'm willing to meet for that.  Sounds like catherineD is in Fremont.  Is ther a coffeeshop with wifi that way?
17:14:19 <zehicle_at_dell> f2f would help move things forward.
17:14:22 <rockyg> We can skyp in others and get it nailed enough to prototype.
17:14:43 <zehicle_at_dell> can you coordinate w/ davidlenwell
17:15:00 <zehicle_at_dell> we don't have critical mass to go further on this one
17:15:10 <zehicle_at_dell> agree?
17:15:23 <zehicle_at_dell> for this meeting (not in general )
17:15:35 <rockyg> Yes.
17:15:40 <zehicle_at_dell> #topic TCUP
17:15:41 <catherineD> rockyg: IBM site is in south bay (near Morgan Hill.  We can drive to Fremont
17:15:52 <rockyg> We definitely need to brainstorm.
17:16:12 <catherineD> rockyg: +1
17:16:25 <zehicle_at_dell> I follwed Praveen's instructions and got tempest to run from TCUP.  I'm pending an edit to his pull but it's looking good
17:16:28 <rockyg> We can meet at the endo of a BART line, or in RWC by the Caltrain.
17:16:41 <rockyg> Yes!
17:17:00 <rockyg> catherineD: where do you live?
17:17:16 <rockyg> my coffee cup is empty :(
17:17:34 <praveen_dell> I am hoping to edit that document tommorrow as I was busy with my regular stuff
17:17:44 <catherineD> I live in Cupertino but do not mind driving
17:17:53 <davidlenwell> Hi guys
17:17:58 <davidlenwell> sorry .. train delay
17:18:12 <rockyg> Train BART?
17:18:21 <zehicle_at_dell> #topic Refstack dev focus short term
17:18:24 <rockyg> ^or^
17:18:35 <davidlenwell> yes bart
17:19:07 <zehicle_at_dell> I pushed the topic back to Refstack Dev
17:19:09 <rockyg> so, f2f soonest.  We can mee at the southernmost end of a BART line.
17:19:23 <davidlenwell> lets get back to that in a minute
17:19:34 <zehicle_at_dell> pending topics are: TCUP, Summit, managed JSON artifacts, Defcore
17:19:51 <rockyg> Goes with the topic.  Nailing down the API spec
17:20:26 <rockyg> API upload spec.
17:21:00 <davidlenwell> I'm working on a spec for that based on the one I already submitted and the notes given by raymond
17:21:05 <davidlenwell> it will be updated today
17:21:17 <davidlenwell> we do not have to have a f2f about the api spec
17:21:32 <rockyg> k
17:21:53 <davidlenwell> also I had a chat with catherine earlier this week.. we agreed on several more phone calls leading up to them releaseing a spec that allows api->api communication of results
17:22:08 <davidlenwell> meaning local refstack can talk to remote refstack .. i would very much like to stay on track with that plan
17:22:23 <davidlenwell> with an agreed upon spec we can finish that code in a day or two
17:22:31 <davidlenwell> without the spec it could eat weeks of rework
17:22:47 <davidlenwell> please please please please write the spec .. I am available for as many phone calls that are needed to get there
17:22:55 <rockyg> kool
17:23:17 <davidlenwell> if we fly by the seat of our pants leading up to the summit we are going to be caught with our pants down
17:23:19 <rockyg> Who owns that spec?  Catherine or Raymond?
17:23:24 <davidlenwell> either one
17:23:33 <davidlenwell> catherineD: is who I spoke to about it
17:24:08 <rockyg> Let's get a date locked down for submitting to Gerrit.  catherineD?
17:24:27 <catherineD> This is the one with Sync and authorization .... We need to discuss with David about authorization protocol ..
17:24:40 <catherineD> which method for authorization etc ...
17:24:46 <davidlenwell> catherineD: lets do a phone call this afternoon to discuss the details of that
17:25:01 <catherineD> That is one of the topic for f2f ..
17:25:10 <catherineD> ok
17:25:22 <davidlenwell> we'll circle back to face to face in a few
17:26:05 <zehicle_at_dell> added to agenda :)
17:26:09 <davidlenwell> thank you
17:26:25 <davidlenwell> okay lets move on to goals for the sumit .. then we'll talk about the summit session
17:26:31 <zehicle_at_dell> #topic summit
17:26:46 <davidlenwell> so among our goals for the summit..
17:26:55 <davidlenwell> * distributable tcup
17:27:02 <davidlenwell> * live results gathering
17:27:29 <davidlenwell> * finalized release version specifications
17:28:04 <davidlenwell> * remote api to central api sync
17:28:28 <davidlenwell> live results gathering is basically finished.. however there was a question about how to flag things as finished ..
17:28:39 <davidlenwell> RaymondWong: to answer your questions about that..
17:29:23 <davidlenwell> I believe we can simply use an error response code / time out to tell if a test fialed for soem reason .. and the act of submitting results back to the api should indicate its finished status
17:30:04 <davidlenwell> so we need distributable tcup and the remote api sync
17:30:20 <davidlenwell> if catherine can finish the spec for that by monday I think we are in good shape ..
17:30:33 <davidlenwell> if we need a f2f to make that happen .. we'll discuss it ..
17:30:58 <davidlenwell> today I am going to work with joshua and zehicle_at_delluntil we have tcup working as it should
17:31:09 <RaymondWong> I think we can do a phone call for that... we have some questions about what to sync and how to authenticate.
17:31:11 <zehicle_at_dell> +1
17:31:16 <davidlenwell> joshua is also going to help a lot with how we are hosting the website and displaying data
17:31:35 <davidlenwell> So I think we can meet the goals I've layed out above
17:31:47 <davidlenwell> before the summit .. but we have to stay orginized and keep on it
17:31:53 <davidlenwell> I say that for my self as much as anyone else ..
17:32:08 <davidlenwell> okay .. so moving along to the summit session..
17:32:43 <davidlenwell> I will be preparing slides and a 20 minute talk giving us a deep dive into the current state of refstack followed by 5 or so about where we are going from here..
17:32:45 <zehicle_at_dell> #topic summit
17:33:14 <zehicle_at_dell> joshua and I also have a defcore preso about it
17:33:19 <davidlenwell> the primary goal of this session is to recruite a little more outside help
17:33:40 <davidlenwell> gain a little more acceptance by the infra group .. they will be invited
17:34:16 <davidlenwell> while at pycon I had a lot of talks with jim blair about how we can integrate with their framework.. and I believe we are in sync. ..
17:34:31 <davidlenwell> so I want a few of them in the loop as much as possible ..
17:34:44 <davidlenwell> zehicle_at_dell: can you talk about what it is you are presenting to defcore?
17:35:27 <zehicle_at_dell> yy
17:35:36 <zehicle_at_dell> that's a lot more than Refstack
17:35:58 <zehicle_at_dell> the prmary focus is to show community the list of "core capabilities" that we've come up with from the process
17:36:14 <zehicle_at_dell> that list (really the must-pass test list) should be setup as a JSON file in refstack BTW
17:36:20 <rockyg> I would like a QA/Tempest liasson for Refstack, also.  Lots of new things there.
17:36:30 <zehicle_at_dell> but we need the community to look at that list and understand how we got there
17:36:45 <zehicle_at_dell> we have some graphics and examples that help people understand
17:36:55 <davidlenwell> rockyg: lets come back to that
17:36:56 <rockyg> Kewl.  Is the matrix available/will be available?
17:37:08 <zehicle_at_dell> it's the one we were working on
17:37:23 <zehicle_at_dell> we need to get more review on it next week
17:37:29 <zehicle_at_dell> from the tech committees
17:37:42 <rockyg> So, dev's should be able to see the mapping by viewing the two, but maybe not others.
17:37:54 <zehicle_at_dell> yy
17:38:09 <zehicle_at_dell> I'm going to break it into sub specialities for review
17:38:17 <rockyg> good
17:38:28 <zehicle_at_dell> but first I need more explanation of scoring capablities so they don't freak out about the work
17:38:35 <zehicle_at_dell> which also has made great progress
17:38:41 <davidlenwell> for now we'll populate the json list in refstack with every single test
17:38:52 <zehicle_at_dell> all of that leads back to TCUP because we want to collect data
17:39:02 <zehicle_at_dell> davidlenwell, yes.  that's the plan
17:39:10 <zehicle_at_dell> we need to work on a JSON schema for it
17:39:17 <zehicle_at_dell> that's another topic I wanted for the agenda
17:39:19 <davidlenwell> zehicle_at_dell: we've already basically done that
17:39:24 <zehicle_at_dell> :)
17:39:35 <davidlenwell> months ago catherineD RaymondWong and tedchang sat in a room and talked it out
17:39:38 <zehicle_at_dell> that enough about defcore at the summit?
17:39:39 <davidlenwell> we should have notes
17:39:45 <davidlenwell> yeah .. im happy with that
17:39:50 <davidlenwell> whats the next topic ?
17:39:56 <zehicle_at_dell> let's talk about the json then
17:40:02 <zehicle_at_dell> #topic json
17:40:29 <zehicle_at_dell> the driver test team also needs a place to store json info
17:40:36 <zehicle_at_dell> that's high stakes (like required tests)
17:40:53 <zehicle_at_dell> I'd suggested that github/refstack is a good place for that since it's managed
17:41:01 <davidlenwell> indeed
17:41:04 <zehicle_at_dell> but we likely need an agreed directory for it
17:41:19 <davidlenwell> zehicle_at_dell: thats easy enough
17:41:26 <zehicle_at_dell> yy
17:41:33 <zehicle_at_dell> just want to get naming right
17:41:44 <zehicle_at_dell> and it would make sense to put the place holders there
17:41:48 <rockyg> and a rst file with instructions
17:41:52 <zehicle_at_dell> driver test team has the files (I believe)
17:42:04 <davidlenwell> zehicle_at_dell: since we are sharing with another team .. lets write it up in a spec
17:42:16 <davidlenwell> that way its clear
17:42:17 <zehicle_at_dell> we also need to document the control process - since the review cycle needs to be a bit more transparent
17:42:17 <rockyg> +1
17:42:34 <davidlenwell> zehicle_at_dell: can you elaborate on that please?
17:42:44 <zehicle_at_dell> meaning, a vendor rep would need to +1 on changes to their driver
17:43:00 <zehicle_at_dell> and a defcore committee member (Josh or I likely) wouldk have to +1 changes to the must-pass list
17:43:20 <zehicle_at_dell> so, the review of those files is specialized
17:43:24 <davidlenwell> well adding you guys to the gate should be easy enough
17:43:32 <zehicle_at_dell> we don't have to be gate
17:43:40 <zehicle_at_dell> in fact, I'd rather not
17:43:41 <davidlenwell> you want +approve
17:43:44 <zehicle_at_dell> right
17:43:44 <davidlenwell> not +2
17:44:04 <davidlenwell> I'll have to figure out if we can do that
17:44:08 <zehicle_at_dell> thanks
17:44:12 <davidlenwell> we might need to eventually move things to another repo
17:44:28 <rockyg> Likely, once things get more formal.
17:44:30 <zehicle_at_dell> perhaps - I think it's OK if we don't have all the bells
17:44:50 <zehicle_at_dell> as long as we agree on the process we can use transparency to make sure it's legit
17:44:55 <rockyg> +1  Lots of business stuff gets invloved later on.
17:45:02 <davidlenwell> for now I can agree to not +2 +a something unles syou or josh have +1'd it
17:45:13 <zehicle_at_dell> that's what I was looking for
17:45:17 <davidlenwell> in that folder
17:45:21 <davidlenwell> or file
17:45:30 <davidlenwell> as far as the rest of the code goes .. things will proceed as normal
17:45:33 <rockyg> +1;)
17:45:35 <zehicle_at_dell> +1
17:45:44 <zehicle_at_dell> name of the folder?
17:46:10 <rockyg> thirdparty
17:46:13 <zehicle_at_dell> suggestions?  /reference /truth /requirements
17:46:28 <davidlenwell> /defcore/requirements.json ?
17:46:41 <zehicle_at_dell> not quite
17:46:46 <zehicle_at_dell> I'm OK w/ defcore
17:47:11 <davidlenwell> zehicle_at_dell: can you write up the spec for the format of the json?
17:47:20 <RaymondWong> it may need to be versioned, havana vs icehouse...
17:47:42 <zehicle_at_dell> was thiking /defcore/must-pass.json and /defcore/drivers.json
17:47:47 <davidlenwell> RaymondWong: probably yes
17:47:54 <zehicle_at_dell> RaymondWong, yes OR needs to be in the json format
17:48:03 <zehicle_at_dell> probably better to have a file per version
17:48:08 <davidlenwell> zehicle_at_dell: "/defcore/havana/must-pass.json
17:48:13 <zehicle_at_dell> +1
17:48:13 <rockyg> +1
17:48:44 <davidlenwell> oh .. use underscores in file names not - signs
17:48:56 <zehicle_at_dell> davidlenwell, I'd rather defer to you or others on the json schema because it really drives the web page
17:49:31 <davidlenwell> RaymondWong: I think you would be well suited to write this spec
17:49:34 <zehicle_at_dell> let's keep it simpler.... /defcore/havana/coretests.json
17:49:56 <rockyg> yes!
17:50:26 <RaymondWong> who is going to do the scoring part of code? that will depend on the json, and may have requirements for that.
17:50:37 <zehicle_at_dell> RaymondWong, that's Defcore's job
17:50:58 <zehicle_at_dell> we've got a list of tests that are required for core
17:51:12 <zehicle_at_dell> ...um... will have.  not done yet ;)
17:51:23 <davidlenwell> RaymondWong: we just need to set up the format for the data
17:51:27 <RaymondWong> i meaning implementing the code logic in refstack to compare the result with the requirement to give a pass/fail score on the web page.
17:51:42 <davidlenwell> RaymondWong: you are getting ahead of your self
17:52:00 <zehicle_at_dell> agree, we DO want that but first we need to collect the data
17:52:31 <davidlenwell> okay .. so zehicle_at_dell.. can you give us a really rough idea of what the requirements coming from defcore will look like ?
17:52:47 <zehicle_at_dell> yes
17:52:54 <rockyg> How about if we can just display the results with fails highlighted?
17:52:58 <zehicle_at_dell> basically, it's just a list of tests
17:52:58 <davidlenwell> if you can get that to raymond todayish
17:53:01 <zehicle_at_dell> the test names
17:53:14 <zehicle_at_dell> there's also a concept of capabilities
17:53:31 <zehicle_at_dell> so, we'd also have a mapping of tests into capability groups
17:53:36 <davidlenwell> okay .. lets clear this up ..we do not need to display anything right now
17:53:36 <zehicle_at_dell> likely that should be a JSON file too
17:53:49 <zehicle_at_dell> davidlenwell, correct.  that's future function
17:53:55 <davidlenwell> right now we need to collect data
17:54:15 <zehicle_at_dell> I think we'll need a second JSON file for /defcore/havana/capabilities.json
17:54:19 <davidlenwell> so having this thing in place that parses the data through this json file we are taling abotu is not currently required
17:54:27 <rockyg> OK.  So, where do these json files get stored?
17:54:33 <zehicle_at_dell> but, again, that's not to be linked into the UI yet
17:54:35 <davidlenwell> in the repo rockyg
17:54:39 <rockyg> The results, that is.
17:54:55 <zehicle_at_dell> davidlenwell, correct.  not required yet
17:54:55 <davidlenwell> the results are stored RAW in the database
17:55:14 <davidlenwell> we'll process them later .. catherineDhas done a good job writing tools that allow us to bulk process data
17:56:05 <zehicle_at_dell> david - do you want me to put those files in as placeholders?
17:56:20 <davidlenwell> zehicle_at_dell: yes please
17:56:24 <zehicle_at_dell> kk
17:57:03 <davidlenwell> RaymondWong: we can talk offline about what I expect from a spec for this
17:57:13 <davidlenwell> it is 11 .. so we do need to wrap up in here
17:57:33 <zehicle_at_dell> #topic wrapup
17:57:50 <zehicle_at_dell> I think we'll need to make specific plans for summit next meeting
17:57:56 <rockyg> what's left undiscussed on agenda?
17:58:15 <davidlenwell> I'll be in #refstack all day .. joshua will be working with me .. please jump in and nag for things
17:58:15 <zehicle_at_dell> you got f2f covered?
17:58:27 <catherineD> David and we will ahve a call to discuss authentication ...
17:58:30 <davidlenwell> we're delaying it for now
17:58:38 <davidlenwell> catherineD: lets say at 4pm
17:58:41 <rockyg> works for me.
17:58:47 <catherineD> OK
17:59:28 <rockyg> back over to #refstack for me for just a little bit.
17:59:29 <davidlenwell> okay zehicle_at_dell don't forget to end the meeting here
17:59:34 <zehicle_at_dell> I'm online most of the time - got to drive in for a few meetings but then back.  I'll ping in and out of #refstack
17:59:39 <zehicle_at_dell> #endmeeting