# [EM] Markus: BeatpathWinner algorithm

MIKE OSSIPOFF nkklrp at hotmail.com
Thu Jun 2 22:36:59 PDT 2005

```Russ had said:

>You might also recall that Mike's "beatpathwinner"
>algorithm ...

>It wasn't "Mike's". It was from Steve, and I said so.
>Russ got the algorithm from me, just as I got it from Steve.
>As I said, it wasn't "Mike's" algorithm. I got it from Steve,
>and Russ got it from me.

Markus says:

That isn't quite correct

Actually, it's quite correct.

Markus continues:

I had posted the Floyd algorithm in a
private mail (30 April 2000) to David Catchpole, Blake Cretney,
Steve Eppley, Rob Lanphier, Norman Petry, and you.

And you sent that to us for no particular reason, as a sort of random cold

Or was there perhaps a different reason why you sent it to us--like for
instance because we'd previously mentioned it.

You see the problem: Why send that to us if we hadn't previously mentioned a
similar algorithm?

You continue:

Obviously, you
didn't understand the Floyd algorithm because your implementation
had a runtime of O(NumberOfCandidates^5) while the Floyd algorithm
has a runtime of O(NumberOfCandidates^3).

I'd never heard of the Floyd algorithm at the time that Steve devised his
BeatpathWinner algorithm.

Steve hadn't heard of the Floyd algorithm at the time that he devised his
BeatpathWinner
algorithm.

Soon after, however, maybe in response to Steve's proposed algorithm, you or
someone posted the Floyd algorithm.

It was what Steve had proposed, except that it only made one pass through
the 3-candidate permutations. I assumed that that difference was an error
that you'd made, and that the actual Floyd algorithm was the same as Steve's
BeatpathWinner algorithm.

So I began calling Steve's BeatpathWinner algorithm "the Floyd algorithm".
Then, when I found out that the Floyd algorithm has a difference, because it
only goes through the 3-candidate permuations once, something made possible
by rearranging the indices in an expression in the center of the loops, I
posted that I'd just found out that the Floyd algorithm is different from
Steve's BeatpathWinner algorithm. From then on, I didn't call Steve's
BeatpathWinner algorithm the Floyd algorithm. I then started just calling it
the BeatpathWinner algorithm.

Markus we've been all over this about 50 times, and that's just counting the
last time we discussed it (endlessly, with you repeating everything again
and again).

Markus continues:

But the fact that you
mistakenly called your implementation "Floyd algorithm" showed
that you read my mail and tried to implement that algorithm that
was described in my mail.

Wrong. I initially did not call Steve's BeatpathWinner algorithm "the Floyd
algorithm". I only began calling it that after you or someone posted the
Floyd algorithm, probably in response to Steve's posting of his algorithm.

Later I called it the Floyd algorithm, for the reason stated above, until an
EM posting pointed out the difference between Steve's algorithm and the
Floyd algorithm.

And no, it was not my implementation. It was Steve's implementation.

Steve devised and described the algorithm before he or I had heard of the
Floyd algorithm.

To other list members:

Last time Markus got started on this subject, he posted for about a month
dozens of times. That's what Markus does. He comes up with an incorrect
chronology, and then endlessly repeats it. He's going to do so again this
time too.

This time I won't reply to more repetition of it. This is my last reply to
Markus on the topic of the BeatpathWinner algorithm, the Floyd algorilthm,
and when I started calling Steve's algorithm the Floyd algorithm, and when I
stopped calling Steve's algorithm the Floyd algorithm, and Markus's apparent
claim that Steve got the idea from him. I've answered Markus's statements
(again) here, and there's no need to keep doing so for the next month.
Answering one time is just right.

So Markus, repeat your same statements again, and you'll have the last word.

Mike Ossipoff

_________________________________________________________________