On Mon, Nov 27, 2000 at 01:16:30AM -0500, Raul Miller wrote:
> On Mon, Nov 27, 2000 at 11:48:05AM +1000, Anthony Towns wrote:
> > But I still don't see how you can read that into the constitution, which
> > simply talks about `strictly more ballots [preferring] A to B'. It seems
> > completely straightforward to read that as meaning `count the number of
> > individual ballots which prefer A to B, count the number that prefer B
> > to A, and compare them. if one's strictly greater than (ie, not equal to)
> > the other, that one dominates the other'.
> And I don't see how you can treat a circular tie as a strict preference.
You're conflating the individual preferences expressed on individual
ballots, and the collective preferences expressed by considering all
the ballots.
Each individual ballot, and each individual voters, expresses a strict
preference amongst all the options that are included in the vote.
However, collectively, there may be no strict preference amongst multiple
options, either due to an exact tie between two options (where neither
dominates the other), or due to a cycle of domination (where A dominates
B, B dominates C, and C dominates A).
But, under any reasonable interpretation I can see, A.6(2) only uses
individual preferences to determine whether one option dominates another.
(This matches, say, the definition of "pairwise-victory" given in
http://www.fortunecity.com/meltingpot/harrow/124/defn.html)
> However, I do understand that that's your interpretation, so I suppose
> the question is: given that this is your interpretation, what do you
> want to do about it?
Have it clearly and fairly handled by the vote counting process.
> Would you like to amend A.6 to make unambiguously clear that A does not
> dominate B if they're both a part of a circular tie?
>
> Would you like to amend the constitution in some other fashion?
> [If so, what's your criteria for rejecting my interpretation as not
> even desirable?]
I reject your interpretation because it doesn't seem to me to match
what the constitution says, not because it doesn't come up with the
right result. It's a similar thing to Darren's "modifying the social
contract will require a 3:1 supermajority". IMO, that's the correct way
to modify the social contract (and hopefully Darren, Manoj and Branden
will eventually work out amongst themselves why we're not voting on how
the social contract can be modified at some point), but just asserting
that that's how things work, because after all the constitution didn't
explicitly forbid it, seems wrong.
> > Even if you change that to "more ballots strictly prefer A to B", you
> > still end up with the same thing. Ditto "strictly more ballots strictly
> > prefer A to B". Ditto "more ballots prefer A to B".
> This is why I have a problem with your interpretation -- you're
> claiming that "strictly" has no meaning, in the constitution.
In maths, the term "strict" when used with comparators (especially
subset, but also less than or greater than) is used to explicitly remove
the equality case. If something's astrict subset of something else,
they're not the same. If something's strictly less than something else,
they're not the same. If something's strictly more than something else,
they're not the same. Using the term without "strict" can be amiguous on
that score (at least in maths), and it would clearly be bad if two tied
options were to be considered to dominate each other (since both would
then be removed, rather than the tie being broken by a casting vote
later). So my assumption is that the term's there to remove any chance
of ambiguity, not to completely change the way everything else is done.
> > It doesn't say "if the ballots collectively express a strict
> > preference for A over B" or something similar, which might be able to
> > be reasonably interpreted how you seem to want to.
> Well, it certainly doesn't make sense to treat the ballots in any sense
> other than collectively -- after all, we're trying to determine their
> collective impact, not their individual impact.
How votes are treated collectively (that is, what they're collective
preference is) is a difficult matter to fairly determine. There are
dozens of ways of doing which result in different winners, all of which
are "fair" in some particular sense or other. Exactly which method of
interpreting a "vote" is to be used is an important matter, it's not
something that should just be intuited by the vote counter.
I would much rather the constitution clearly and precisely describe how
the vote should be undertaken, rather than expecting the secretary to
correctly judge obscure ambiguities or invent interpretations so we can
get any result at all.
> However, I guess I understand that you don't want to see it that way,
> which leaves us with the question: now what?
We should work out exactly how we want votes to happen, and we should clearly
and explicitly document that in the constitution.
Cheers,
aj
--
Anthony Towns <aj@humbug.org.au> <http://azure.humbug.org.au/~aj/>
I don't speak for anyone save myself. GPG signed mail preferred.
``Thanks to all avid pokers out there''
-- linux.conf.au, 17-20 January 2001
Attachment:
pgptwKyavxTbb.pgp
Description: PGP signature