[EM] 2nd Matt reply--12/20/03

Markus Schulze markus.schulze at alumni.tu-berlin.de
Sun Dec 21 03:31:02 PST 2003


Dear Mike,

when I wrote (on 15 Dec 2003) that you called your implementation
"Floyd algorithm" there was no reason for me to believe that you
have changed your opinion recently. Craig Carey claimed that my
implementation of the Floyd algorithm doesn't work (presumably
because it makes only one pass through the triple-loop and
presumably because Craig believed because of the while-loop
in your implementation that the Floyd algorithm doesn't find
the strongest paths in a single pass through the triple-loop).
Therefore, I stressed that your implementation is not the Floyd
algorithm and that your algorithm doesn't find the strongest
paths in a single pass. But when one considers the possible
short cuts in that order that has been proposed by Floyd then
a single pass is sufficient.

******

You wrote (21 Dec 2003):
> That algorithm wasn't written as an attempt to write the Floyd
> algorithm. It was written by people who had never heard of the
> Floyd algorithm. It was written to accomplish the pupose that
> it accomlishes. Only later, due to Markus's earlier post of
> something he called the Floyd algorithm (though it was about
> strongest paths rather than shortest paths), did I hear of the
> Floyd algorithm.

Even though you write that you have never heard of the Floyd
algorithm, it is a matter of fact that you called your algorithm
"Floyd algorithm" e.g. in the source code of your Python program
(http://electionmethods.org/CondorcetSSD.py). Therefore, when I
stress that your implementation is not the Floyd algorithm then
this is a feasible observation and not an "attack". If you hadn't
mistakenly written that your implementation is the Floyd algorithm,
then there wouldn't have been a reason for me to stress that your
implementation is not the Floyd algorithm.

******

You wrote (21 Dec 2003):
> I've been futily trying to explain to Markus that I'm not claiming
> that Steve's algorithm is the Floyd algorithm. Yes, I admit that
> that effort has been futile.

When I wrote (on 15 Dec 2003) that you called your implementation
"Floyd algorithm" there was no reason for me to believe that you
have changed your opinion recently. In one of your recent mails,
you admitted that you mistakenly believed Eppley's algorithm to
be Floyd's algorithm and you admitted that you mistakenly called
Eppley's algorithm "Floyd's algorithm". Therefore, my criticism
was feasible and you admitted this. Therefore, there is absolutely
no justification for your insulting mails.

******

You wrote (21 Dec 2003):
> Can you name one false thing I said about Markus? I merely pointed
> out the falsity of his continually-repeated statement that I claim
> that Steve's stongest beatpaths algorithm is the Floyd algorithm.
> That's really the only issue in the discussion.

I don't claim that you continue to claim that Eppley's algorithm
is Floyd's algorithm. However, when I pointed (on 15 Dec 2003) to
the fact that you mistakenly called your implementation "Floyd
algorithm" this observation was true.

Markus Schulze



More information about the Election-Methods mailing list