>> Only the last vote is taken. (I check for duplicate
>> names and take only the last vote from each given
>> person, ignoring any previous votes.)
kuwahara> Well, the problem is that you assume 1) names
kuwahara> are unique 2) people use their own names
kuwahara> whenever they cast their votes.
Yes, of course. For #1, hopefully people don't call
themselves "Sam Iam" in one case and "Jenny Doe" in another
:-). As long as the variants are no worse than "M Hucka"
and "Mike Hucka", I can figure it out.
For #2, the poll specifically asks for people's names. If
people can't be bothered to write their names, then we have
kuwahara> Even if these assumptions hold, with this
kuwahara> approach, anyone can vote
But I have a list of subscribers to sbml-discuss, and can
check against the list to determine if given voters appear
to be reasonably legit. So if someone outside of
sbml-discuss votes, there's a high probability of being able
to spot it.
If a name appears multiple times (which hasn't happened so
far -- no one's revoted), I can also check one other piece
of information to correlate separate entries by that name in
the voting record. For now, I'd like to not disclose what
that piece of information is, for the little bit of security
it adds to the scheme. It's not usuable in all cases, but
it's a small addition.
kuwahara> I understand this is an interim thing (thank you
kuwahara> for pointing me to your SBML specification
kuwahara> process proposal) But still, I think a better
kuwahara> means should be taken.
Yes, I'd like that too. But it's worth keeping in mind that
we're not voting in presidential elections here. The stakes
aren't that high, and I don't think we're at high risk of
voter fraud, so I don't think we need to spend a lot of time
and resources creating a complicated high-security voting
system. It's hard enough getting people to vote at all!
(But to be fair, the main difficulty is it also requires
people to spend time thinking about sometimes complicated
and muddy technical issues, and everyone's busy, so it's a
lot to ask.)
But thanks for your comments nonetheless. I do worry about
the issues you bring up too, and if we could find better
schemes, we should consider them.