[Linux-aus] AGM Announcement

[Stewart Smith]

Anthony Towns said:
> Any chance the Jan 6th/7th stuff could be separated out slightly more;
> so that there's time to read the committee reports before the chance to
> raise any questions is over? Would committee reports coming out
> sometime between Dec 25th and Jan 2nd, be too much to ask?

possibly... there will be beer involved around that period. Although I
don't think we should prevent reports from going up sooner if they're
ready.
> (Recall that last
> time, Michael Davies got nominated for VP on the last day of
> nominations or so, and then declined the nomination; the above might
> mean people would already have voted for him by the time that was
> cleared up)

With a preferential system, this won't be a problem. A pref system also
deals with the situation where someone stands for more than one position.

> I was thinking of making up a voting system that lets you vote for
> things like "I want as many different states represented as possible",
> and "I don't want more than one person from OzLabs" and "I'd prefer it
> if Stuart were VP, but if he doesn't get that I'd like to vote for him
> for President", but the "right" way to do that is kind of complicated
> and confusing, and I decided it probably wasn't worth the effort
> anyway.

I think a preferential system will give the best balance for this.


> So what are the ballots going to look like? They can't be the same as
> we have in real-life -- ie, one vote for P, VP, S, T, three names for
> ordinary members -- because unlike at LCA itself, you can't decide
> "well, if Pia doesn't get to be P, I want to vote for her as VP, but
> otherwise I'll vote for Stuart to be VP". Presumably then we want to
> have preferences, so that we write out:
>
> 	President:
> 		1. Alice
> 		2. Bob
> 		3. Carol
> 		4. Dave
> 	Vice-President:
> 		1. Alice
> 		2. Carol
> 		3. Fred
> 	Treasurer:
> 		1. Elice
> 		2. Dave
> 	Secretary:
> 		1. Geraldine
> 	Ordinary Members:
> 		1. ...
> 		2. ...
>
> And then just iterate through, eliminating people who've already won
> something?

Yeap... that's my thoughts.

> Using the same system we have in federal elections would seem to work
> okay. Is that what'll be used?

I hope so :)

Basically, the committee has to agree to it. I think they will :)


> Do we want votes to be anonymous?

I don't think this is reasonably possible. We'll be storing a membership
number alongside the vote. This allows people to go back and edit their
vote (e.g. if someone drops out for whatever reason, or they prove to be a
real asshole) as well as preventing someone from voting twice.
> Do we want to have people be able to verify the results of the vote? Do
> we want people to be able to directly verify their vote has been
> correctly counted?

We could release the list of votes (without member numbers) so that people
could perform their own checks. Would this be worthwhile? Is there any
point? The voting code will be open source (GPL), so people will be able
to verify its integrity.
Although they will be taking our word for it that it's what's actually
running there (although those with access to the machine will be able to
actually do that... if you trust them :)
> Do we want it to be impossible to prove how you've voted to someone
> else?
>
> (Debian does all but the last for the leadership election; the last two
> are mutually exclusive, I believe)
>
> I guess "Dec 10th, run a test of the voting system software" was
> accidently missed from the timeline? :)

With luck :) The more people testing it the better. I'll be trying to
co-ordinate public tests of this.
-- 
Stewart Smith (stewart at flamingspork.com)
http://www.flamingspork.com