Path: utzoo!utgpu!news-server.csri.toronto.edu!cs.utexas.edu!sun-barr!apple!rutgers!cmcl2!kramden.acf.nyu.edu!brnstnd
From: brnstnd@kramden.acf.nyu.edu (Dan Bernstein)
Newsgroups: comp.lang.misc
Subject: Re: A very brief history of optimal sorting methods
Message-ID: <26029:Nov1919:29:5090@kramden.acf.nyu.edu>
Date: 19 Nov 90 19:29:50 GMT
References: <5293:Nov1518:36:0490@kramden.acf.nyu.edu> <12926:Nov1604:11:1490@kramden.acf.nyu.edu> <6177@lanl.gov>
Organization: IR
Lines: 53

Little technical content. (This is the ``liar or fool'' subthread.)

In article <6177@lanl.gov> ttw@lanl.gov (Tony Warnock) writes:
>     Dan has called me a liar and a fool,

No, only one or the other.

> With respect to the
>     former, I shall only quote Dan's own words.

I observe that none of the quotes (except the one you took entirely out
of context---see below) match your original claims. Hence I will settle,
with all due respect, upon calling you a liar. I also observe that the
quotes match what I said in my initial response, and so you have
contributed absolutely nothing to this discussion.

> "The number of the partition chosen takes at most a few bits to encode."

You and Jim Giles seem to believe that I want to compile infinitely many
copies of my code. As I have said repeatedly (including in the original
presentation of the method), the number of partitions should be limited
to some fixed number. If that number is 20, as I once suggested, then
each partition takes at most 5 bits to encode. If it is 2, in the
all-out strategy, then each partition takes 1 bit to encode.

You accused me of saying that an exponential (or whatever the word was)
quantity could be encoded in a few bits; if the above quote is your only
justification, as I conjectured in the article you're responding to,
then you are a fool. I prefer to think that you're just lying and trying
to weasel your way out of the hole you've dug.

> "Luckily, on the other side of the fence, we programmers still know that
> sorting is linear."

You ripped this out of the context, in which I had already made it clear
that I was talking about time per byte.

> "Fortran does not have separate compilation like C."

Yes. Observe the ``like C.'' You accused me of saying that Fortran does
not have separate compilation. You lied.

>  
>                                                         - Dan Bernstein

Was the blank line intentional there?

> "I thank God for not making me a computer scientist."
>                                                         - Dan Bernstein

And I still do.

---Dan