The Election Committee announces the second Steering Committee election

Announcements
64

Right. Worst case, two candidates with same CoI end up popular enough to otherwise be elected, in which case we may have to look into this, at which point maybe we’ll end up with enough experience for this not to be a concern in future cases.

Now, in the case of same CoI, ‘hope only one turns out popular’ is pretty valid, as we may not otherwise need to have found a technical solution.

Once we do solve this however, I think there would be another potential use-case, given a constitutional change: preserving proportional representation (even if by approximation) in the event delegates step down early, by passing on such seats in a way better respecting the delegate’s voter preferences than the current ‘the seat remains vacant until elections’ or ‘pick the next candidate in the original ranking (without taking into account preferences of the withdrawn candidate)’.

3 Likes
65

We have a technical solution for same-CoI actually elected, just a bit annoying to perform and then write/publish a report about: drop the less popular one and recount. The way proportional-ish systems work it means that those voters who would have given the votes to elect the dropped candidate would now support their next-best choices (and in case they are aligned there, it can be pretty orderly).

This can theoretically lead to some weird effects on the edges, in the sense that it might also affect which of the two similar candidates with very different CoIs and positions gets elected (if it was a coin-toss of five votes among a hundred and a half, not much balance shift is needed to flip it, and sometimes changes propagate). Basically this is what makes it most annoying!

If we are willing to have an official forked/customisable tallying implementation, not just go with the available ones and cross-check them against each other, we might be able to figure out a way to recount the previous election with one candidate dropped and still-sitting SC members definitely counting as elected.

However, I would not advise trying this as an official project before everyone agrees that permanently-running contribution-based-access (voting — but also NixCon ticket priority queue, and maybe other stuff?) register is official and working. Just to make sure we can, coordination-wise, do the thing that is simpler to implement and only needs a majority to be official, before trying to get constitutional status to a technically involved project.

3 Likes
66

FWIW, I’m very familiar with how this works in practice in other contexts, and I think it’s great :slightly_smiling_face:

1 Like
67

ah darn I missed my chance to put my hat in, maybe next time then.

68

So, EC has converged both on methods in principle and on implementation in practice of sending, via official infra (so, no third parties asked to do something with the emails as emails), reminders to voters not yet registered — including magic links to «confirm this email»

The rate-limited mass-mailing is currently rolling.

Note that cheap methods (maintainers-list / reuse of the 2024 work) leave some emails unclear (see link).

There are GitHub emails that are recognised in commits as authorship, but not shown via API. In principle, one could check which of the commit emails from PRs from recognised commits look reasonable. In practice it is manual work and I teach all day tomorrow (and I am not at liberty to discuss how exactly the workdays of my EC colleagues are creating availability limitations, but we all know that life happens — and don’t worry, the availability is not zero). However, checking 60 or 70 links to good recent commits in the evening is more feasible, so if someone wants to increase the chances the missing reminders get sent soon, you could help with looking at PR search / commit lists / git log and send us a list of usernames with matching GH commit links.

11 Likes
69

So.

The first (and probably the only large one) batch of magic-link registrations applied.

Of course I managed to mess up two things, one is being slightly more crawable than planned (fixed soon-ish, then checked logs — it’s unambiguous that in any case no crawler obtained enough information to forge consent; and it was only surprisingly well-behaved crawlers); and the second one, spending like a day with wrong content-type.

Worry not, as OpaVote is well-behaved, and is mentioned in the first line of the letter, clicking the link is intent enough for us, even if you did not see the sub-kilobyte «thank you» static webpage.

We are just above 480 actively registered voters now.

(Help with tracking down the missing emails via commit emails will be gratefully accepted then used)

7 Likes
70

The voting has been opened, OpaVote emails should have been sent (and if not, it is now worth investigating).

We should be able to handle some late registrations. But it is manual and laggy (as mentioned before).

PS. I find A note about STV elections a good effort in explaining some corner cases of STV

PPS. Voting will be stopped manually as soon as feasible after the expiry of 2025-11-01 Anywhere on Earth (i.e. after 23:59:59 UTC-12)

11 Likes
71

Up to

Some numbers on this, as I enjoy stats:
Last year, we had 841 eligible voters. This year, we had 1010 eligible voters, giving us a 20.09% increase in eligible voters.

Last year, 450 people actually ended up voting. While we don’t have actual voting numbers yet, we do have the registration numbers. There are currently 535 people registered to vote[1].

Of course, not everyone who registered will vote - and there will be people who register late. But if we pretend that these two factors perfectly cancel each other out, we get a 18.88% increase - very close to the 20.09% increase in eligible voters.

Finally, thanks to the EC team for sending out emails anyways, and allowing late registration. I still think mandatory voter registration was a mistake - but thanks to their work, I don’t think it’ll lead to a meaningful decrease in turnout.

  1. I cloned the repo, deleted the header line, and ran cut -f3 -d, eligible.csv | grep . | wc -l. We use grep . to delete any empty entries, for people who haven’t submitted their emails. ↩︎

7 Likes
73

Just a reminder that voting will end in a bit less than 24 hours. So if you have not voted yet and are eligible, please do so! You should be good as long as Anywhere on Earth - Time.is still says it’s Nov 1st.

6 Likes
74
4 Likes