Does the Nov 16 - Nov 30 voting timeframe still work?

[ambv (Łukasz Langa)]

@ambv

I would like to open a new topic to choose a new date for the Big Vote.

I received a question about this sentence so to clear things up for posterity: the vote is on between Nov 16 and Nov 30 as originally planned. The poll on the voting model was swift and with a clear winner so moving the Big Vote is unnecessary.

[njs (Nathaniel J. Smith)]

@ambv

the vote is on between Nov 16 and Nov 30 as originally planned

This is news to me, and the three-day notice is kind of freaking me out. My todo list still at least one more round of tweaks I need to apply to 8016, and two discussion topics that need to be opened: one to find out whether there are enough people interested in serving on council to make those proposal viable, and a pre-vote survey to spur discussion. (I was hoping to post all these today.)

In the original PEP 8001, if discussion spilled over into the voting period that was OK, because people could always change their votes. But in this thread it was decided to switch to a system where it’s impossible for people to change their mind after they vote. There seemed to be general agreement there that this was only workable if we had a pre-poll survey…

[tim.one (Tim Peters)]

EDIT

I’m closing this poll, in favor of the newer one asking PEP authors how much time they need.

@ambv

the vote is on between Nov 16 and Nov 30 as originally planned.

@njs

This is news to me,

The relevant PEP (8001) didn’t change the date, and hasn’t been touched for over a week now. There was no discussion of changing the date here either.

@njs

In the original PEP 8001, if discussion spilled over into the voting period that was OK, because people could always change their votes. But in this thread it was decided to switch to a system where it’s impossible for people to change their mind after they vote.

As mentioned before, people definitely need to be aware that they can’t change their vote. But the amount of time they have to consider their vote hasn’t changed - they’d be well advised to vote toward the end of the voting window, unless they’re sure they won’t change their mind.

@njs

There seemed to be general agreement there that this was only workable if we had a pre-poll survey…

Whereas I read it as that people didn’t want to bother with a Discourse poll. The quite strong (to me) evidence is that nobody in fact did bother to create one, despite that it’s so easy to do and was last suggested a week ago.

I don’t really care when the vote starts or ends, but there has been strong consensus that we need to get it done. The announced dates haven’t changed, and have been spelled out in the controlling PEP all along. It will always be “too soon” for someone, yes?

That said … which voting window(s) are acceptable to people?

• 16 Nov - 30 Nov (current)
• 23 Nov - 7 Nov (delay a week)
• 30 Nov - 14 Dec (delay 2 weeks)
• 7 Dec - 21 Dec (delay 3 weeks)
• 14 Dec - 28 Dec (delay 4 weeks)
• spill into next year

[njs (Nathaniel J. Smith)]

@tim.one

The relevant PEP (8001) didn’t change the date, and hasn’t been touched for over a week now. There was no discussion of changing the date here either.

Łukasz’s message started by quoting his comment, in this thread, about changing the date…and obviously the date wasn’t real while the discussion was ongoing and we didn’t know when it would be finished. I was assuming that the process would be something like, at some point someone proposes consensus on the discussion having finished and the PEP being ready to move out of “Draft” status, and that would be the obvious point to check in on the timing. (Note that as of right now it’s still marked as “Draft”.)

@tim.one

I read it as that people didn’t want to bother with a Discourse poll. The quite strong (to me) evidence is that nobody in fact did bother to create one, despite that it’s so easy to do and was last suggested a week ago.

Well, now you have even stronger evidence the other way

I’m not sure what to say in the poll. I want to get this done soon too, but I don’t like cutting off substantive, productive discussion. Obviously 3 days is too soon for that. Maybe 10 days is enough, maybe not, I don’t know? And IIRC people also wanted to avoid having the polling period overlap with the holidays, so even a one week delay might be equivalent to pushing it into next year, I’m not sure.

[tim.one (Tim Peters)]

@njs

I’m not sure what to say in the poll.

Discourse polls appear to support only Plurality (pick one) and Approval (pick as many as you like) directly, and Approval is more revealing. So list the PEP numbers and ask people to check off all the ones they approve of.

@njs

Maybe 10 days is enough, maybe not, I don’t know? And IIRC people also wanted to avoid having the polling period overlap with the holidays, so even a one week delay might be equivalent to pushing it into next year, I’m not sure.

It already overlaps with Thanksgiving in the US. I don’t see any point in people arguing about it, so just vote in the poll I already created for the voting window:

https://discuss.python.org/t/python-governance-electoral-system/290/114?u=tim.one

[njs (Nathaniel J. Smith)]

@njs

I’m not sure what to say in the poll.

I meant I don’t know what to put in your poll

[tim.one (Tim Peters)]

@njs

I meant I don’t know what to put in your poll

Haha! I missed that .

But It’s not meant to be an intellectual puzzle: which date ranges are you personally OK with? Check all that apply. It’s not asking what you believe other people “should” also be OK with - just what works (or doesn’t) for you.

[malemburg (Marc-André Lemburg)]

Shouldn’t the question be: are all the relevant PEPs ready to be voted on ?

If you can answer yes to that question, the next question is: “How long do people need to review all of them, ask questions, etc. ?” and then you pick a week for voting after this period has ended.

I can’t answer the first question, but since all PEPs are fairly complex esp. when it comes to details, I’d say we need to give people at least two weeks for reviews and questions.

[ambv (Łukasz Langa)]

There’s a few things I need to respond to and a general comment so I’ll use fancy headers to signify what I’m focusing on.

Time for reviews and questions

@malemburg

I’d say we need to give people at least two weeks for reviews and questions.

Right. PEPs published early gave people more than a month for reviews and questions. This happened amply on Discourse and the mailing list already. Interested parties could always come in and ask questions. Others decided not to engage and that is their choice.

Pushing the vote later

@njs

I want to get this done soon too, but I don’t like cutting off substantive, productive discussion. Obviously 3 days is too soon for that. Maybe 10 days is enough, maybe not, I don’t know? And IIRC people also wanted to avoid having the polling period overlap with the holidays, so even a one week delay might be equivalent to pushing it into next year, I’m not sure.

Exactly right: pushing it by even “one week” will likely mean we are pushing the vote to 2019.

Inability to change your vote: WHAT?

@njs

But in this thread it was decided to switch to a system where it’s impossible for people to change their mind after they vote.

Wait, what? Why? I don’t see this decided anywhere. The original PEP 8001 specified explicitly that people are free to change their mind for the period of the vote. To change this we would need to have an explicit discussion about why this is an issue, there was none.

If you mean “ Allow voters to write in new choices.” then I understood this as the inability for people to add new options to the poll. This is of course uncontroversial.

General comment: I am troubled by the ensuing chaos.

The first version of PEP 8001 was published on October 15th. It stated, among other details, a date for the vote far enough in advance. In fact, this date was based directly on @Mariatta’s original proposal from August 1. There were no suggestions for any other dates and any concern voiced about inadequate time for reviews and questions.

The first batch of governance PEP placeholders was posted to the repository in late August by @barry. Before anyone started work on those, we spent significant time researching how third-party projects are governed and wrote down summaries in PEP 8002.

Then came the core sprint in Redmond when the initial crop of PEPs got volunteers to actually author it. @steve.dower quickly penned his own new proposal, followed by @jackjansen who was not at the sprint. @vstinner was not happy with any existing proposal and decided to post a late addition on October 6th. Mariatta’s original timeline for PEP authorship and review was well known and understood, as proven by Victor’s plea to wait for his PEP since there was a non-disputed deadline for original PEP submissions of October 1st that was later extended by a week.

That means PEPs 8010 - 8015 were evolving for a good 6 weeks with ample time for reviews and discussion. In this setup, PEP 8016 (announced two weeks ago!) should not have even been considered but I’ve stayed quiet about this because I didn’t want to hinder what seemed like a healthy discussion. But in exchange I expected the authors of PEP 8016 to understand that they should not hold everybody else hostage due to their late submission.

I hope the ordeal with “putting the PEP 8001 back into draft status” to switch from IRV to Condorcet (and from GitHub votes to CIVS) will work out. Based on Tim’s comments above, looks like provided the voters are of light build and relatively sedentary we might even end up with a Condorcet winner the first time round!

My biggest hope though is that we won’t see any more last minute disruptions.

[njs (Nathaniel J. Smith)]

@ambv

Wait, what? Why? I don’t see this decided anywhere. The original PEP 8001 specified explicitly that people are free to change their mind for the period of the vote. To change this we would need to have an explicit discussion about why this is an issue, there was none.

If you mean “ Allow voters to write in new choices.” then I understood this as the inability for people to add new options to the poll. This is of course uncontroversial.

No, that’s not what I mean :-). I mean that PEP 8001, the one that’s currently in the PEP’s repo, has been revised to make it so that you cannot change your mind after you cast your vote. (Or, well, you can change your mind, but if you do then sucks to be you, because you can’t change your vote.) You probably want to read this thread:

PEP 8001: Public or Private Ballots?

I too would like it better if people could change votes, as often as they like, before the election ends. Against that, CIVS has battle-tested software to get everything else right. So, if we use CIVS, that’s a tradeoff: people need to be aware that they cannot change their vote. And so they’d be well advised to delay voting until near the end of the voting window. A good point: CIVS does not close the election on its own. Our election supervisor has to close it “by hand”. So if a slew …

So uh, I don’t want to put words in your mouth, but putting that fact together with what you wrote, it sounds like you’re saying that PEP 8001 requires further discussion?

@ambv

there was a non-disputed deadline for original PEP submissions of October 1st that was later extended by a week

“Non-disputed” feels a bit revisionist? 1, 2, 3, 4, 5, 6

@ambv

Exactly right: pushing it by even “one week” will likely mean we are pushing the vote to 2019.

I mean, I know it’s no-one’s favorite outcome, but if the end result of all this is that the vote schedule slips once and things are resolved in January, then that’s not the end of the world? Even if we pick one of the proposals that requires a followup vote and that takes a month, then we’ll still be done by the end of Februrary, well ahead of Guido’s deadline of PyCon (which is May 1-9). And everyone who’s been pushing for voting ASAP will still get to feel satisfied that their work kept things moving well enough to keep us down to a single schedule slip :-).

[ambv (Łukasz Langa)]

@njs

So uh, I don’t want to put words in your mouth, but putting that fact together with what you wrote, it sounds like you’re saying that PEP 8001 requires further discussion?

No, I put the rationale behind not being able to change your vote later explicitly in the PEP body, as well as the warning suggested by @tim.one for people to hold off with casting ballots until they are absolutely certain. Also added your bunch as co-authors which was overdue, sorry!

I do think that both Condorcet (through its unclear “is there even a winner?” nature that @tim.one explored above) and CIVS (through the inability to change your mind) make this election rather fragile. This increases the likelihood we’ll have to do this all over again. The original PEP’s goal was to avoid that and now that’s out of the window. I’m not going to veto anything but I’m worried.

@njs

“Non-disputed” feels a bit revisionist? 1, 2, 3, 4, 5, 6

How dare you spoil my argument with FACTS!

Sorry if I sounded dismissive, the way the discussion went then made me misremember those.

@njs

Guido’s deadline of PyCon

Maybe I’m nit picking here but Guido’s deadline was “well before PyCon”.

Do PEP authors need more time?

After thinking about this the entire day, I have to agree with @malemburg that the main audience to ask whether Nov 16 - Nov 30 is realistic is the PEP 801x authors. I created a poll for them to specify if they are on track: PEP 801x authors, are you on track for the vote between Nov 16 - Nov 30?

I also e-mailed all personally in case they are in the Discourse-sceptic crowd.

[Mariatta]

I also think that we should keep the original date of Nov 16.

In addition to what Lukasz had said, the private repo where we collect email address for voting clearly states
“” the General election date as Nov 16:

Missing emails for Governance Election on 2018-11-16 #1

I do admit some fault in this. I intended to send out reminders two weeks ahead of voting and one week ahead of voting, but I got too occupied with other things… I’m sorry.

But still, the Nov 16 date is not new news, it’s been brought up since August, and during the sprint meeting, (you were there too @njs) we talked about the date and everyone in the room heard and agreed to it.

Maybe I’m nit picking here but Guido’s deadline was “well before PyCon”.

I also understood it as “well before PyCon” as in long long time before PyCon (May 1, 2019)

[brettcannon]

FYI I forked this date conversation to its own topic as the other topic was already quite long and this seemed like a rather focused topic.

[brettcannon]

@njs

I was assuming that the process would be something like, at some point someone proposes consensus on the discussion having finished and the PEP being ready to move out of “Draft” status, and that would be the obvious point to check in on the timing. (Note that as of right now it’s still marked as “Draft”.)

I think the status of PEP 8001 doesn’t signal anything as there’s no clear consensus on who gets to make the call it’s finished anyway (since we have no consensus on anything beyond we need a governance model someday ) . I’ve been going on the assumption that we would mark it final on Nov 15 when the vote starts since we are done talking about voting at that point.

[barry (Barry Warsaw)]

Do we have a definitive list of eligible voters? If not, then opening the voting in 2 days seems ambitious.

[Mariatta]

My understanding is this is the voters list. Core devs have been asked to provide their email address there. If they’re still not in the list, perhaps they just don’t intend to vote?

[barry (Barry Warsaw)]

Thanks. I just want to make sure that we’ve used every means at our disposal (e.g python-committers, python-dev, etc.) to reach potential eligible voters. It would be really bad if some core dev missed an announcement until it was too late.

[barry (Barry Warsaw)]

@Mariatta

My understanding is this is the voters list

That list looks incomplete. I’ve seen another list where warsaw is on it, but it’s not on that list. Are there more than one list floating around?

[pitrou (Antoine Pitrou)]

Actually, the voting list is this file. As you can see, several active developers don’t have e-mails in this file (e.g. vadmium, rhettinger, markshannon…).

[Mariatta]

They were asked to provide their email address here, but have not responded yet.