20:01:11 <ttx> #startmeeting tc
20:01:12 <openstack> Meeting started Tue Aug 20 20:01:11 2013 UTC and is due to finish in 60 minutes.  The chair is ttx. Information about MeetBot at http://wiki.debian.org/MeetBot.
20:01:13 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
20:01:15 <openstack> The meeting name has been set to 'tc'
20:01:17 <ttx> Our agenda:
20:01:25 <ttx> #link https://wiki.openstack.org/wiki/Governance/TechnicalCommittee
20:01:35 <ttx> Hopefully will be a short one. The pool awaits
20:01:44 <ttx> #topic Motion to start using Gerrit for TC votes
20:01:51 <ttx> #link http://lists.openstack.org/pipermail/openstack-dev/2013-August/013164.html
20:02:04 <ttx> In summary, mordred proposed that we use Gerrit to track the various iterations over a motion words, and the recording of final approval/rejection votes
20:02:16 <ttx> Then the merged motions would end up in some repository where they would be used as reference
20:02:30 <ttx> mordred: does that summarize it well ?
20:02:38 <mordred> yes
20:02:43 <ttx> Personally I think that would solve some confusion, etherpads and in-channel pastes
20:02:48 <mordred> probably goes without saying that if we approve that...
20:02:54 <ttx> As well as provide an automated record of what we voted on and how we voted on that.
20:02:58 <jd__> o/
20:03:00 <mordred> we'd need to go back through the previous things we've done, make a repo with those things in it
20:03:04 <notmyname> here
20:03:11 <mordred> so that new things we do can be additions to that
20:03:18 <ttx> mordred: sounds a lot less great now that you mention it
20:03:24 <mordred> ttx: :)
20:03:29 <gabrielhurley> +1 to having a repo of important documents
20:03:35 <mordred> "automated record of what we voted on and how we voted on that" is the main driving impetus
20:03:36 <markmc> ttx, I think he volunteered :)
20:03:44 <dolphm> mordred: while that sounds nice, why do you think its necessary?
20:04:05 <dolphm> mordred: but if you're volunteering, then i won't ask why
20:04:05 <ttx> markmc: lgtm
20:04:10 <mordred> dolphm: we've been making decisions in IRC for 3 years now, and it's getting tough to know what the actual current set of policy is for people
20:04:17 <markmc> gabrielhurley, yeah - we're not great at recording our motions in the wiki after the fact
20:04:25 * gabrielhurley despises wikis
20:04:32 <annegentle> gabrielhurley: high5
20:04:39 <annegentle> markmc: noooo
20:04:41 <mordred> so, obviously we can have this auto-publish on commit and whatnot
20:04:43 <ttx> mordred: most TC-era motions should be found in the meeting minutes I post after meetings
20:04:44 * wikis despise gabrielhurley
20:04:45 <gabrielhurley> wikis: I despise you
20:04:52 <mordred> and I'm not volunteering myself, but I do volunteer resources to have it done
20:04:59 <ttx> but pre-TC motions might just be forgotten or inappropriate
20:05:04 <dolphm> mordred: ttx: i assume TC 'reviews' would be public for any ATC to comment on?
20:05:12 <mordred> ttx: well, pre-tc motions lead to the tc charter
20:05:13 <ttx> given future changes
20:05:17 <mordred> dolphm: YES!
20:05:28 <ttx> mordred: ok, so post-TC motions at least
20:05:29 <mordred> dolphm: so, I thik that "proposing a motion" should be "propose a commit to the tc repo"
20:05:30 <markwash> can we just add the TC meeting notes?
20:05:31 <markwash> summary sounds complex and potentially legalistically confusing
20:05:41 <ttx> mordred: then we'll see if we need t orevive some pre-TC ones
20:05:41 <markwash> yes, legalistically
20:05:56 <dolphm> mordred: so, any ATC can propose changes - correct?
20:05:56 <ttx> mordred: So my only concern is that it should not encourage us to bypass discussion (ML and meeting)
20:06:04 <ttx> dolphm: yes
20:06:11 <vishy> o/
20:06:13 <ttx> mordred: So I would add two constraints:
20:06:18 <ttx> 1. Prefer the ML to discuss the whole concept and alternative solutions
20:06:19 <russellb> ttx: i thought that too, but i'm not sure it's much different than what we have now
20:06:26 <russellb> between ML and then discussion/vote in this meeting
20:06:30 <ttx> 2. Require that the motion stays up for review until at least one TC meeting has occurred to discuss it (we can even vote during that meeting)
20:06:45 <mordred> ttx: that works for me
20:06:45 <ttx> then I think with these safeguards on, we are golden
20:06:59 <ttx> NB: I checked and I don't think we need to change the TC charter to pass that
20:07:00 <markwash> did we decide how to represent rejected decisions?
20:07:10 <ttx> The only thing that is no longer totally applicable is proxying, but we can keep proxies and handle them "manually" using comments
20:07:11 <markwash> or do we skip that?
20:07:25 <mordred> markwash: as abandoned changes? :)
20:07:49 <ttx> markwash: I don't think we need to track rejected motions, further than the review link
20:08:07 <mikal_laptop> Sorry I'm late, DSL problems
20:08:09 <markmc> ttx, remind me how the voting works again? who has approve rights?
20:08:12 <mordred> ttx: where is the current TC charter?
20:08:17 <markwash> how does the gerrit discussion get archived?
20:08:23 <notmyname> ttx: markwash: rejections, or the reasons for them, can be just as important as things that are approved
20:08:23 <mordred> markmc: I was suggesting we give the TC chair APRV rights
20:08:30 <markmc> mordred, sounds good
20:08:44 <ttx> mordred: http://wiki.openstack.org/Governance/Foundation/TechnicalCommittee
20:08:47 <dolphm> ttx: so even with gerrit, we'd wouldn't use "asynchronous" voting? (we'd be voting during the course of a regular meeting?)
20:08:55 <markmc> this will encourage people to properly document proposals upfront too
20:09:10 <ttx> markmc: TC members have +2, Chair has APRV to tally votes
20:09:12 <vishy> do we need proxies if we are going to have longer to vote?
20:09:25 <annegentle> markmc: not sure it really will. Etherpad/wiki is quite easy to use for proposals.
20:09:27 * mordred doesn't think so - but doesn't feel strongly about it
20:09:55 <markmc> annegentle, we'll nitpick the hell out of these proposals in gerrit, like we do with everything else in gerrit
20:09:56 <ttx> vishy: I think we can keep them, if only to avoid having to introduce a special motion to chnage the charter
20:09:57 * mordred was talking about proxies
20:09:59 <annegentle> I also think we should have the votes in the meeting, because isn't attendance at a certain number of meeting a requirement for serving?
20:10:06 <ttx> vishy: but I would expect their usage to drop
20:10:24 <vishy> sounds like there is confusion about whether votes will be async
20:10:39 <vishy> based on dolphm and annegentle 's comments
20:10:43 <dolphm> vishy: ++
20:10:45 <markwash> for some reason I really like the idea of having records of official-ish TC activity in a git repo. . which is why I'm interested in recording rejections and tracking discussion somehow
20:10:47 <annegentle> I would support Gerrit for wording/policies, but voting at meetings
20:10:57 <ttx> vishy: the idea would be async votes, but you don't close the vote until the end of the next TC meeting
20:11:18 <ttx> annegentle: we can vote using Gerrit during the meeting
20:11:30 <ttx> I actually inhtend to cast my +2/-2 during the meeting
20:11:30 <russellb> ok, so proxies are probably less relevant then
20:11:31 <annegentle> ttx: true, we could
20:11:44 <ttx> and use +1/-1 until the meeting is held
20:12:02 <mikal_laptop> Well, if you can't make the meeting surely you can +2 early?
20:12:11 <mikal_laptop> Thus avoiding a proxy?
20:12:16 <russellb> right
20:12:21 <ttx> but if people are confident they will never change their minds, they can use +2/-2 alright
20:12:33 <ttx> mikal_laptop: I like to be convinced DURING the meeting
20:12:47 * mordred likes to hold on to pre-conceived notions in the face of reason
20:12:50 <mikal_laptop> Yeah, cool. So the +2 during the meeting bit is more of a guideline than a rule. Like the pirate code.
20:12:58 <mordred> mikal_laptop: ++
20:13:11 <ttx> right.
20:13:21 <dolphm> mordred: can ATC's vote +1/-1 to express support (or not) of a motion? (assuming TC members can cast +2/-2)
20:13:30 <mordred> dolphm: yeah. I don't see why not
20:13:31 <ttx> dolphm: yes
20:13:40 <mordred> it would be an interesting way to record that sort of thing
20:13:57 <mordred> also, I think we'd set the merge mode in gerrit to be cherry pick
20:14:03 <mordred> which means votes will get aded to the commit message
20:14:16 <markwash> mordred: oh cool
20:14:17 <ttx> mikal: we would close the vote if we reach enough +2/-2 during the meeting. Otherwise we'd just let the vote continue over the nex tweek async
20:14:19 <mordred> (seeing as how I do not expect 100 patches to this repo per week)
20:14:22 <markwash> mordred: +1 to that
20:14:26 <notmyname> maybe a silly question, but does gerrit have any issues with having dozens of +/-2s or hundreds of +/-1s on a particular patch?
20:14:34 <ttx> until we get enough +2/-2
20:14:42 <mordred> notmyname: it shouldn't - although it'll be interesting to see :)
20:14:50 <mordred> notmyname: the votes are all in a db table
20:14:52 <markmc> that would be fun
20:14:55 <markmc> 100s of +1s
20:14:56 <dolphm> notmyname: i like the way you think
20:15:04 <mordred> the UI might look terrible at that point
20:15:11 <mordred> s/might/will/
20:15:30 <ttx> identifying issues where generic people care enough to cast a vote is useful, too
20:15:35 <annegentle> I think asynch with ATC votes is a good way to get input from the ATC crowd, better than IRC.
20:15:59 <dolphm> is the TC repo named? ;)
20:16:28 <ttx> Let me know when you're ready to vote... one of the last ones we'll have maybe :)
20:16:28 <markmc> trove is a nice name for a repo of documents
20:16:39 <jd__> humpf
20:16:41 <markwash> markmc: lol
20:16:45 <ttx> markmc: "quantum" is nice name too
20:16:46 <annegentle> markmc: bwah ha
20:16:51 <mikal_laptop> I'm ready to vote
20:16:54 <notmyname> I'm imagining a situation where something is open for a bit to get ATC feedback and is a hot issue. it could very easily get lots of participation and attention. that's one of the whole charter goals for openstack (open governance)
20:16:58 <jd__> would it need incubation too?
20:17:00 <mikal_laptop> Seagate is a great name for a networking product
20:17:16 <markmc> notmyname, yes, that would be a great side-effect of the change
20:17:23 <ttx> notmyname: +1
20:17:31 <mordred> notmyname: +1
20:17:39 <russellb> +2 APRV
20:17:40 <ttx> notmyname: we'd use common sense in not closing an issue too fast
20:17:45 <notmyname> just wait until techcrunch or gigaom is promoting at TC vote ;-)
20:17:51 <ttx> notmyname: hence my recommendation to not cast +2/-2 too early
20:17:52 <mordred> maybe we'll get fancy and convince the board to do this too ...
20:18:24 <dolphm> #ready
20:18:28 <gabrielhurley> #set
20:18:32 <ttx> let's do this
20:18:32 <notmyname> so with that said, I consider markwash's comment about tracking discussion and rejections importatn too
20:18:35 <russellb> #gogogo
20:18:48 <gabrielhurley> russellb: thanks. someone needed to finish it. ;-)
20:18:54 <notmyname> and if we have a tool that does that, use it instead of blueprints /anothertopic
20:19:18 <ttx> #startvote Use Gerrit in the near future to track motions and record TC votes? yes, no, abstain
20:19:19 <openstack> Begin voting on: Use Gerrit in the near future to track motions and record TC votes? Valid vote options are yes, no, abstain.
20:19:20 <openstack> Vote using '#vote OPTION'. Only your last vote counts.
20:19:21 <markmc> #vote yes
20:19:25 <markmcclain> #vote yes
20:19:25 <russellb> #vote yes
20:19:27 <jd__> #vote yes
20:19:27 <ttx> #vote yes
20:19:29 <markwash> notmyname: we might just have to set up a record of gerrit history somewhere
20:19:29 <dolphm> #vote yes
20:19:30 <shardy> #vote yes
20:19:34 <mikal_laptop> #vote yes
20:19:35 <vishy> #vote yes
20:19:36 <notmyname> #vote yes
20:19:42 <markwash> #vote yes
20:19:44 <ttx> 30 more seconds
20:19:50 <markmc> markwash, is gerrit not a record of gerrit history?
20:19:50 <mordred> #vote yes
20:20:03 <markwash> markmc, not one I can access offline as easily, or backup in as many places
20:20:13 <ttx> #endvote
20:20:14 <openstack> Voted on "Use Gerrit in the near future to track motions and record TC votes?" Results are
20:20:15 <openstack> yes (12): markmc, ttx, notmyname, vishy, shardy, jd__, russellb, markwash, mordred, dolphm, mikal_laptop, markmcclain
20:20:22 <ttx> motion approved
20:20:28 <russellb> yay
20:20:33 <mordred> markwash: I _believe_ newer gerrit is moving to storing discussion in the git repo
20:20:35 <ttx> mordred: need anything frmo us/me to get this going ?
20:20:35 <notmyname> markmc: and I'd imagine that rejections, by being called rejections (or abandoned) have the connotation of being disposable
20:20:38 <mordred> markwash: but I have not verified that
20:20:50 <jd__> markwash: we need to make Gerrit use Git repository for tracking votes :)
20:20:52 <mordred> ttx: nah - I'll sync up with you to talk about getting the repo set up and populated
20:21:01 <ttx> mordred: I'll be around
20:21:06 <mordred> markwash, jd__: I believe upstream gerrit is already working on that
20:21:11 <markwash> jd__: turtles and elephants all the way down!
20:21:14 <ttx> #topic ATC exception for Jaromir Coufal and Liz Blanchard
20:21:15 <mordred> they want to make everything stored in git and nothing in a database
20:21:16 <jd__> mordred: awesome :)
20:21:22 <ttx> #link http://lists.openstack.org/pipermail/openstack-tc/2013-August/000336.html
20:21:35 <ttx> We have a provision in our charter for exceptionally granting ATC status to technical contributors when their contribution does not end up as a commit in a git repo
20:21:50 <ttx> gabrielhurley nominated Jaromir Coufal and Liz Blanchard for their work on Horizon UX
20:21:51 <dolphm> (we probably need to do this way more often)
20:21:58 <ttx> gabrielhurley: could you do a quick praise ?
20:22:01 <gabrielhurley> sure
20:22:16 <ttx> mordred: (sidenote) we could actually maintain the list of "extra ATCs" in that gerrit-for-motions thing, together with associated project and expiration date
20:22:20 <ttx> mordred: so we could programmatically use it when we compute ATC lists
20:22:46 <ttx> (like for election setup purposes)
20:22:58 <russellb> clever
20:23:06 <mikal_laptop> ttx: although perhaps set an approval expiry in gerrit...
20:23:10 <gabrielhurley> They've both been leading the efforts to build a UX community for OpenStack, contributing to discussions both in the Horizon community and on the ML for OpenStack in general. Moreover they're actively engaged in blueprint work, helping to get people to the right design decisions. I think that makes them active technical contributors even if they never commit a line to github/gerrit.
20:23:15 <mikal_laptop> ttx: and then just renew by tweaking the config
20:23:30 <dolphm> mikal_laptop: base it on the date from git blame?
20:23:39 <russellb> gabrielhurley: sounds like a great use of this rule then
20:23:44 <gabrielhurley> I thought so :-)
20:23:47 <mikal_laptop> dolphm: renewing is harder then, right?
20:23:53 <gabrielhurley> I would be sad if they weren't at the design summit
20:24:04 <dolphm> mikal_laptop: ah, hmm..
20:24:05 <markwash> sounds like an easy vote!
20:24:05 <mikal_laptop> Anyway, yes. These people sound like ATCs to me.
20:24:07 * jd__ don't want a sad gabrielhurley
20:24:10 <gabrielhurley> heh
20:24:15 <ttx> Do we need two separate votes or one will do ?
20:24:17 <russellb> me either!
20:24:22 <mikal_laptop> One
20:24:23 <russellb> sounds like one should do
20:24:26 <gabrielhurley> one is fine with me
20:24:30 <jd__> one's fine
20:24:38 <ttx> Any question before we #startvote ?
20:24:59 <dolphm> #ready
20:25:10 <russellb> #set
20:25:18 <gabrielhurley> #set
20:25:19 <ttx> Note: Charter says the exception is valid one year (two elections)
20:25:21 <notmyname> IMO we should vote separately
20:25:24 <markwash> #set
20:25:29 <ttx> (like any commit)
20:25:34 <mordred> ttx: yes to above sidenote
20:26:02 <ttx> notmyname: except that nothing in what was presented to me let me distinguish one from the other
20:26:09 <dolphm> ttx: not implying that notmyname would vote differently for each candidate, but if anyone wants to, then a separate vote is a must
20:26:17 <notmyname> ttx: good point
20:26:21 <gabrielhurley> two votes is also fine with me
20:26:28 <ttx> I'm fine with two votes. Let's do this
20:26:33 <markwash> three votes!
20:26:34 <gabrielhurley> #gogogo
20:26:37 <gabrielhurley> ;-)
20:26:41 <ttx> #startvote Grant Jaromir Coufal exceptional ATC status? yes, no, abstain
20:26:42 <openstack> Begin voting on: Grant Jaromir Coufal exceptional ATC status? Valid vote options are yes, no, abstain.
20:26:43 <openstack> Vote using '#vote OPTION'. Only your last vote counts.
20:26:44 <russellb> #vote yes
20:26:45 <gabrielhurley> #vote yes
20:26:45 <markmcclain> #vote yes
20:26:47 <markmc> #vote yes
20:26:47 <annegentle> #vote yes
20:26:48 <mikal_laptop> #vote yes
20:26:48 <dolphm> #vote yes
20:26:48 <notmyname> dolphm: I think 2 votes would be good as a precedent and consistency with future votes
20:26:50 <markwash> #vote yes
20:26:50 <ttx> (no we won't vote on who should go first)
20:26:52 <notmyname> #vote yes
20:26:55 <shardy> #vote yes
20:26:57 <jd__> #vote yes
20:26:58 <ttx> #vote yes
20:26:58 <dolphm> notmyname: agree
20:27:02 <vishy> #vote yes
20:27:06 <ttx> 30 more seconds
20:27:20 <dolphm> ttx: we've had time to internet-stalk both candidates
20:27:32 <holms> #vote yes
20:27:35 <ttx> #endvote
20:27:36 <openstack> Voted on "Grant Jaromir Coufal exceptional ATC status?" Results are
20:27:37 <openstack> yes (14): markmc, ttx, notmyname, vishy, annegentle, jd__, shardy, russellb, holms, markwash, gabrielhurley, dolphm, mikal_laptop, markmcclain
20:27:56 <ttx> #startvote Grant Liz Blanchard exceptional ATC status? yes, no, abstain
20:27:57 <openstack> Begin voting on: Grant Liz Blanchard exceptional ATC status? Valid vote options are yes, no, abstain.
20:27:58 <openstack> Vote using '#vote OPTION'. Only your last vote counts.
20:28:00 <gabrielhurley> #vote yes
20:28:01 <russellb> #vote yes
20:28:02 <markmcclain> #vote yes
20:28:02 <dolphm> #vote yes
20:28:02 <markwash> #vote yes
20:28:02 <mikal_laptop> #vote yes
20:28:03 <annegentle> #vote yes
20:28:04 <markmc> #vote yes
20:28:06 <shardy> #vote yes
20:28:06 <holms> #vote yes
20:28:11 <notmyname> #vote yes
20:28:11 <ttx> isn't voting fun ?
20:28:13 <vishy> #vote yes
20:28:14 <ttx> #vote yes
20:28:16 <jd__> #vote yes
20:28:17 <gabrielhurley> wheeeee!
20:28:19 <ttx> 30 more seconds
20:28:36 <jd__> will be funnier with Gerrit!
20:28:41 <ttx> #endvote
20:28:42 <openstack> Voted on "Grant Liz Blanchard exceptional ATC status?" Results are
20:28:43 <openstack> yes (14): markmc, ttx, notmyname, vishy, annegentle, jd__, shardy, russellb, holms, markwash, gabrielhurley, dolphm, mikal_laptop, markmcclain
20:28:45 <jd__> *click* *click* *click*
20:28:46 <dolphm> /vote for gerrit
20:28:53 <gabrielhurley> great. thanks folks!
20:29:04 <ttx> notmyname: t occurred to me that under the new use-gerrit-to-vote, abstain will look a lot like not voting at all
20:29:13 <ttx> making it a lot less fun
20:29:17 <notmyname> ttx: we have a +0
20:29:35 <ttx> but those look like comments most of the time.
20:29:38 <dolphm> ttx: i use +0 regularly as an intentional "this is my feedback but i'm abstaining from a +1 or -1"
20:29:41 <markwash> gabrielhurley: its great to bring in people who help out the community without necessarily being code committers
20:29:47 <gabrielhurley> :-)
20:29:54 <ttx> notmyname: I guess it would work if you're careful with it
20:30:04 <ttx> #info Both motions approved
20:30:13 <ttx> #topic Open discussion
20:30:18 <ttx> Anything, anyone ?
20:30:55 <notmyname> ttx: ya, seems like an abstain should be explored
20:31:20 <ttx> "After using git/gerrit for code and infrastructure, OpenStack decides to use it for governance too"
20:31:37 <ttx> notmyname: I think Gerrit 2.6 has something along those lines configurable
20:31:51 <ttx> if I remember what jeblair said in the thread correctly
20:32:36 <markwash> regarding the uploading history to the git repo
20:32:49 <markwash> srsly can we just put the meeting logs in a directory in that repo?
20:33:01 <markwash> I don't think we immediately need a cleaned up summary of all the things
20:33:16 <markwash> *old* meeting logs, that is
20:33:22 <ttx> markwash: if we just do TC-era, it should be easy
20:33:27 <markwash> okiedokie
20:33:28 <dolphm> markwash: ah to establish history?
20:34:08 <markwash> a scifi dream world where you monitor your government by doing git pull :-)
20:34:15 <gabrielhurley> lol
20:34:25 <gabrielhurley> can I --rebase?
20:34:26 <gabrielhurley> or --amend?
20:34:29 <dolphm> lol
20:34:32 <ttx> mordred: ideally the TC charter wiki page should be a readonly mirror of a version we'll put in the governance repo
20:34:32 <jd__> lol
20:34:33 <markwash> push -f!
20:34:37 <mordred> ttx: yes
20:34:56 <gabrielhurley> ttx: +1 to that!
20:35:01 <mordred> ttx: or a redirect page to a published page on openstack.org/ even
20:35:08 <markwash> I'm pretty sure huxley would have a field day with my silliness. . maybe he already did
20:35:13 <jeblair> the governance repo can be a sphinx doc and published on openstack.org, then the wiki can link there
20:35:16 <ttx> mordred: whatever
20:35:45 <ttx> ok, we'll flesh out implementation in the next weeks
20:35:58 <ttx> Any last minute comment before we close ?
20:36:17 <dolphm> can we use markdown?
20:36:34 <ttx> dolphm: anything wrong with XML ?
20:36:43 <dolphm> ttx: OOH I DIDN'T KNOW THAT WAS AN OPTION
20:36:56 <ttx> XML is enterprise. Always an option
20:37:11 <mikal_laptop> Let's port the TC to Java
20:37:15 <holms> xml is enteprise :)))))
20:37:18 <jd__> considering how OpenStack is going entreprise
20:37:26 <ttx> heh
20:37:27 <ttx> #endmeeting