Friday, 2026-08-14

gouthamrdansmith: bauzas noonedeadpunk: I'm planning to get the 2027.1 runtiem update merged by Monday.. was wondering if you wanted to look/comment in the meantime:  https://review.opendev.org/c/openstack/governance/+/99692505:35
opendevreviewMerged openstack/election master: Remove Venus from 2027.1 election  https://review.opendev.org/c/openstack/election/+/100091010:51
opendevreviewMerged openstack/election master: Adding Arnaud Morin candidacy for Mistral  https://review.opendev.org/c/openstack/election/+/100091610:59
dansmithgouthamr:  yep, sorry, got it.. AFAIK bauzas is out for a couple of weeks13:35
opendevreviewJeremy Stanley proposed openstack/election master: Sort owners CLI opts by short alias  https://review.opendev.org/c/openstack/election/+/100097615:19
fungitc-members: while implementing the new changes to election tooling for https://governance.openstack.org/tc/reference/charter.html#voters-for-tc-seats-ac i'm trying to follow the letter of the charter, but there is a minor discrepancy...15:43
fungiin particular it talks about merged changes (the prior qualifier), and then says "...who reviewed or approved one of the changes meeting the above criteria (with CR+2, CR-2 or W+1)..."15:43
fungiso as written, any of those votes applied to a change that hasn't merged won't qualify the reviewer as ac15:45
fungiand in particular, a change can't merge with a cr-2 on it, so that case will never occur, leading me to wonder if that interpretation may not be the intended one15:45
fungialso if i applied a cr+2 to a change before the qualifying period but then the change merged during the qualifying period, that would make me ac even though i wasn't actually active during the qualifying period (granted the same goes for code contributors who proposed a revision before the qualifying period which then merged during)15:47
fungianyway, i just want to confirm that what i'm currently implementing, which will match the requirement as written in the charter now, is that for any change merged during the qualifying period, anyone who had a cr+2 or w+1 vote recorded on that change counts as ac15:48
fungiif the next tc wants to amend the charter further to relax that rule so that it counts the set of cr+/-2,w+1 votes applied to any change during the qualifying period rather than only to changes merged during the qualifying period, we can further adjust the election tooling to accommodate that15:51
spotz[m]Yeah I think it could be changed to submitted or proposed17:16
fungialso, due to limitations in gerrit's data model, i can really only effectively get the final votes of users on the last revision of the merged change17:18
fungiit doesn't store prior votes or per-revision values17:19
fungiso if i +2 a change and then it gets revised and i don't +2 again before it merges, that doesn't count17:19
fricklerbut older votes are shown in the UI, so they must be stored somewhere in gerrit?17:20
fungibut in all probability that will be good enough, if someone's getting a ballot or running for ptl without any merged changes, they probably ought to have plenty of core review activity17:20
fungiwhere are older votes shown in the ui, you mean in the comment records?17:21
gouthamrin my reading, this wasn't a contradiction17:21
frickleryes17:21
fungiwe could include those, but it would take days to perform the query iterating over every comment17:21
gouthamrthis would be a really small rare/edge case anyway17:21
gouthamryeah17:22
fricklerhmm, o.k., then maybe finish the implementation first and then we amend the policy to what the gerrit API actually allows17:22
frickleredge cases can still be handled by the extra AC procedure. if someone notices them early enough, that us17:23
frickler*is17:23
gouthamryeah, if the tooling can produce a result17:23
gouthamrthat we can share/audit, this would be easier for someone to flag17:23
fungii've got the implementation for rolls sketched out but am doing a bit of refactoring first so it can be a lot more dry (since it shares a lot of logic with existing data processing which is full of short-circuits and i don't want to double the script runtime)17:25
fungicandidate qualification will be easier since it's just adding a single direct query17:25
gouthamrfungi: have we considered publishing electoral rolls and allowing folks to request exceptions? 17:26
fungiwe could. i'm sort of iffy on sticking a giant list of e-mail addresses on the web (even though they're already public elsewhere), but maybe we could filter it to just their display names or something17:27
spotz[m]gouthamr: What exceptions?17:27
gouthamrno not emails17:27
gouthamrspotz[m]: i'm thinking, if someone's missing on the electorate, we allow them to tell us why they must be included and we make it happen17:28
fungispotz[m]: an edge case where someone isn't a code contributor but did set core review votes only on non-final revisions of changes17:28
spotz[m]So folks who the election officials said weren't qualified?17:28
gouthamryeah, spitballing here for legitimate cases, because we have so many things that can go wrong/missing for you to not get a ballot (or be qualified to run for any governance/office), or get credited for participating in a release17:30
fungitrying to include them in the automation would explode the script runtime by probably at least an order of magnitude, and odds are there is no individual actually in that set to begin with, but further the overlap of that set with people who would actually cast a vote or run for ptl is even smaller still17:31
gouthamryeah17:31
gouthamri understand this pain better now, thanks for trying it out and bringing the concern here.. let's see what we can do, and let the TC/community know how we'd handle this election so we can improve it for next time17:32
spotz[m]I'm just trying to think of what would go on the roll and why. Because we should be telling them if not qualified and they should be getting back to us to ask for the exception. So a roll would be in addition to that process with them still being responsible for asking for the exception17:33
gouthamrtoday we don't do exceptions.. 17:33
fungior rather exceptions were codified as the extra_ac lists17:35
fricklerwell the exception would could do today is extra-ac. but that option ends today. while contributions are counted until next week https://governance.openstack.org/election/#electorate17:35
frickleralso looking at that page I see that my IRC nick has gone missing yet again :-(17:35
gouthamrwe expect you to have fixed up all the qualifying things:17:35
gouthamr- be a code/documentation contributor (i.e., have a change merged) in the past year17:35
gouthamr- be a foundation individual member17:35
gouthamr- be a reviewr that has exercised +2/+W in the past year (brand new) 17:35
gouthamr- or, be added as an extra-ac17:35
gouthamron top of this make sure you have your gerrit ID on your foundation profile, or you'll never get counted17:35
spotz[m]An sometimes we forget we need extra AC, back in the day Tom sent out a request if anyone had any to be added17:36
gouthamrhow did he do this more effectively than us (me/TC, release team, election officials) sending emails to the openstack-discuss list for weeks?17:37
spotz[m]I think for one thing we had separate email lists back in the day so they seemed more targeted to you17:38
gouthamri see17:38
gouthamrhmm, this was one of the things we said we'd figure out with the CEWG.. so keep the ideas flowing :) 17:41
fungitom fifield specifically used to reach out to the various groups that were part of the user committee electorate, not the tc as far as i can remember17:43
fungiback when there were separate tc and uc defined in the foundation bylaws17:43
spotz[m]Yeah I mainly remember it from WoO17:43
clarkbmany election systems ask the voter to register if interested. Automatic registration improves turnout and participation so this effort is good, but what if we publish a roll $time before election and tell people to ask to "register" if they believe they are missing17:45
clarkbhonestly I think if you did that you'd be able to avoid these complicated updates to the tooling entirely since what we're talking about are a small number of corner cases that would be easily verifiable by hand, but harder to capture every election automatically17:45
gouthamr+117:46
spotz[m]I kind of like that because sometimes I hven't realized it until after I didn't get a ballot17:47
fricklerI don't think publishing a list of persons is a good idea, especially today. currently everyone can run the election tooling to generate their own roll locally, but I'm not sure how gerrit would feel if 1k contributors did that at the same time17:51
fricklerso maybe expanding the tooling to only check which lists a specific contributor would be in could be a nice idea17:52
clarkbmaybe I have a different perspective because my government election roll status is public17:52
clarkbI don't personally have an issue with openstack having public rolls as a result. YOu can get far more info from the state I live in17:53
opendevreviewJeremy Stanley proposed openstack/election master: Refactor contributor recording into a new function  https://review.opendev.org/c/openstack/election/+/100099117:54
opendevreviewJeremy Stanley proposed openstack/election master: Reformat lines in record_contributor function  https://review.opendev.org/c/openstack/election/+/100099217:54
clarkbthey will charge you $500 for the whole list though17:54
clarkbbut for anyone trying to abuse the system thats a minimal fee17:54
clarkbThe range of voter roll access worldwide is interesting. You can definitely tell that trying to balance transparency with privacy is something elections offices find different approaches for. The UK has public rolls that you can opt out of but otherwise are available. Australia has public rolls but they can only be viewed in person at elections offices18:11
gouthamrIndia published them publicly, online that anyone can get to iirc18:12
gouthamrbut they're in a form and format that's tedious - maybe they assumed that'll secure them. in any case, having usernames perhaps is a good choice here18:13
gouthamrinstead of divulging emails/names18:13
gouthamrif folks want they can create an account on gerrit and get to those.. 18:14
JayFI'll note that treating contributor emails as PII is a bit silly given we publish them (git repos)18:15
JayFI'm not sure what information election rolls would have more sensitive than (your email + name)18:16
* gouthamr agrees18:17
dansmithwe're talking about just the list of the electorate and not who voted or didn't in a given election right?18:19
gouthamryeah just the electorate.. results/voting details are anonymous (and should stay so)18:20
dansmiththe latter seems highly PII to me, but the former is sort of a consequence of being an open community18:20
dansmithwell, I assume "who you voted for" would always be private but even "did you vote" is not cool to divulge18:21
JayFdansmith: it's a good point; we should protect the secrecy of an "abstain" ballot, even if in 99.9999% of cases it's someone just not paying attention to18:21
dansmithJayF: absolutely.. for sure, "did you vote or not" is highly sensitive IMHO18:21
dansmithin past years when winning openstack elections was a blessing instead of a curse there was definitely concerted effort to get someone elected with in an org18:22
dansmithand abstaining is one way to not participate in that sort of thing, but being visible that you did would be problematic18:22
JayFI was 100000% thinking of that /exact/ scenario18:23
* dansmith nods18:24
gouthamrdo you two collaborate on fridays?18:24
dansmithyes, I -1 everything JayF says on Mon-Thu and then agree on Fridays.. it's all part of my secret plan to remain enigmatic, but now you've found me out18:25
gouthamr:'D 18:26
JayFthe bright sunshine of the weekend at the end of the tunnel18:26
dansmithI also coordinate t-shirts with slawek on Wednesdays18:26
JayFgives us enough recognition of the outside world 18:26
JayFto have a sliver of hope18:26
JayFmaybe things will get better! [hears train horn] oh 💩; MONDAY! RUN18:26
fungidansmith: i don't think the current system (civs) even tells the election officials who has voted, only how many have voted19:34

Generated by irclog2html.py 4.1.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!