> > Minor nit: it's (C,A) which is weakest, as (A,C) is not a proposition.
On Sun, Nov 17, 2002 at 06:31:19PM +0100, Jochen Voss wrote:
> Why not? To quote from your Nov 17 draft:
I changed the definition of proposition between Nov 16 draft
and Nov 17 draft. That above minor nit was in the context of
the Nov 16 draft.
> > > A B C
> > > A - 27 0
> > > B 22 - 0
> > > C 0 0 -
> > > the Schwartz set is { A, C }
> >
> > Hmm.. I forgot to eliminate options with no votes for them.
>
> Be careful here: in this example my program agrees with Anthony
> Towns implementation, both get a tie between A and C. If we
> eliminate options with no votes for them (C in this case) we
> change the result: now the option A becomes the winner.
Yes: A got 27 votes prefering it over B, B got 22 votes prefering it
over A, all preferences involving C have been eliminated.
C should have been excluded from the Schwartz set, because it has
no votes.
> It may be possible, that we want that change. But we should
> know that this is a deviantion from our former implementation.
It's a difference from the former implementation because the former
implementation was wrong.
--
Raul
