PEP 308: Voting Procedures

[Raymond Hettinger]

Laura pointed out that a straight acceptance vote
is biased toward "no change".  The bias arises
because some who want a ternary operator would
still prefer "no change" to certain alternatives.

One way to solve this one is a two pass vote
procedure with the first step selecting the most
popular alternative and then having a straight go
or no-go vote on the result.

This would also be the most helpful for
Guido:  here is the community's most
accepted proposal and here is the
proportion that prefer it to no change at all.

If Aahz wants to be clever, I can image that
he can optimize this into a single vote with
instructions like, "rank order the alternatives
from most acceptable to least acceptable, then
for each one mark whether you prefer it to no
change at all."

Besides removing the built-in bias, it has the
advantage of allowing those who prefer no-change
to still be able express which alternative they find
to be the least unacceptable.


Raymond Hettinger