[EM] Reply to Matt

[MIKE OSSIPOFF]

Message: 3
From: matt at tidalwave.net
Date: Fri, 19 Dec 2003 08:52:15 -0800 (PST)
To: election-methods at electorama.com
Subject: Re: [EM] Reply to Ernie
Reply-To: matt at tidalwave.net

Matt said:

In this case the style of debater and the merit of the argument positively
correlate.  In other words, the debater that is polite, clear and helpful is
correct and the debaters that are impolite, confusing and derogatory are
incorrect.

(farther below, Matt clarifies that he's saying that Markus was the correct 
one)

I reply:

Sorry, but you're mistaken.

The "debate" was about whether or not I claim that Steve's BeatpathWinner 
algorithm is the Floyd algorithm. Since I repeatedly clarified that, now 
that I've been told different, I now am not claiming that Steve's algorithm 
is the Floyd algorithm, Markus was incorrect when he kept repeating that I 
claim that Steve's algorithm is the Floyd algorithm.

That's the only issue that was being debated. On that issue, I was correct, 
and Markus was incorrect, because, during that discussion, I was longer 
claiming that Steve's algorithm is the Floyd algorithm.

I wasn't debating which algorithm is better. I'd be lying if I claimed to 
know if one pass through the 3-candidate permutations can reliably find the 
strongest beatpath beween each ordered pair of candidates, as Markus claims, 
but I'm not debating that or denying it.

Matt continued:

Thank you Markus for your (unsuccesfull) efforts to help Mike improve the
program that his site promotes.

I reply:

Yes, when Markus said that Steve's algorithm doesn't work, Markus was 
unsuccessful in convincing me that it doesn't work. But you're sure that it 
doesn't work, right? :-)   I'm referring to the
BeatpathWinner algorithm that I posted here a few days ago.

Or maybe the improvement that yoiu're referring to is the greater speed of 
the one-pass algorithm, assuming that it relliably works.

Let's suppose that it indeed works, and is faster than Steve's algorithm. 
Does that mean that we should use it to replace Steve's algorithm at the 
website, or that I should start sending it out to people who ask for an 
implementation algorithm. No, not really. Steve's algorilthm _obviously_ 
works. It's obvious that eventually the repeated passes will find the 
strongest beatpath between each orderred pair of candidates.

Even if the one-pass algorithm works, I'm not going to send it out, along 
with a proof for why it works. I prefer to not send something that requies 
me to convince someone and explain to them why it works, when it doesn't at 
first appear that it would.  I'll continue sending out the algorithm that 
obviously works, Steve's algorithm. Sorry, Matt.

Matt continues:

I have seen stubborness before but the
magnititude of this stubborness (over such a trivial yet clearcut issue!)
matches or exceeds the worst I can recall seeing before.

I reply:

What issue? There was no issue about the relative merits of the 2 
algorithms. Markus says that there's a 1-pass algorithm that works, and that 
it's faster. Though I told how I initially didn't believe that that could be 
done, I was _not_ debating that during the discussion. My position has been: 
I don't claim to know if that can work, but it isn't important.

The issue, as I said, was about whether I claim that Steve's program is the 
Floyd algorithm. Yes I agree that that's a trivial issue. And yes, I 
stubbornly maintain that I don't claim that Steve's algorithm is the Floyd 
algorithm :-)   How's that for stubborn?

Or is that you think that I'm stubborn because I don't replace Steve's 
algorithm with the putative one-pass algorithm? I've told you that I prefer 
something whose validity is obvious, something that doesn't require a proof 
to be sent with it, doesn't require as much explaining and convincing. How's 
that for stubborn? If Steve's algorithm runs for a half minute, and the 
one-pass algorithm runs for half a second, is half a minute really a big 
problem?

So you're saying that it's stubborn to not adopt the method that you like 
best?

But, at the time that you stated that I was stubborn, I hadn't yet said that 
I wasn't going to adopt the one-pass algorithm.  All I'd said at that time 
was that I don't know if Markus's one-pass algorithm works. I was being 
honest. I don't know that if it works. I haven't checked Floyd's 1962 proof, 
and I don't even know if that proof really applies to the job of finding 
strongest beatpaths. That isn't stubborn, it's honest.

So you're saying that it's stubborn not to take Markus's word for whatever 
he says, in spite of Markus's long history of mis-statements?

So it's really entirely unclear what you think that I was being stubborn 
about.

Mike Ossipoff

_________________________________________________________________
It’s our best dial-up Internet access offer: 6 months @$9.95/month. Get it 
now!  http://join.msn.com/?page=dept/dialup