[Osmf-talk] Share Alike images

[Matthias Urlichs]

On Di, 2009-12-08 at 08:06 +1100, Elizabeth Dodd wrote:

> I would expect then a run-off poll like in France, we eliminate the
> least popular choices and revote.

Problems: (a) who decides on the magnitude of "choices", (b) it will
probably take longer, (c) whether run-off o not, it can lead to strange
vote results.

Consider these results:
40% B
25% A1
20% A2
15% A3

Now if the A voters would prefer any of the Ax to B, it's fairly obvious
that some A option should win -- but if the result disappoints enough of
them that they don't re-vote, B wins the revote. Not a good thing.

> I do not want preferential voting as in Australia (we are all always
> confused by it) 

What's confusing about ranking a list of options in order of "best,
second best, ..."?


In any case, the Australian IRV method has a couple of _other_ nasty
ranking problems, which are avoided by the Condorcet method (used by the
Debian project, among others).

There's a free voting service at www.cs.cornell.edu/andru/civs.html --
I'd suggest that you give it a shot.