From: "Jan Simon" <>
Newsgroups: comp.soft-sys.matlab
Subject: Re: Derangement: efficient full permutation
Date: Tue, 1 Feb 2011 22:08:03 +0000 (UTC)
Organization: Universit&#228;t Heidelberg
Lines: 14
Message-ID: <iia083$qki$>
References: <ii8ut8$3fb$> <ii90lg$qmd$> <ii91fn$5eo$> <ii920m$s53$> <ii98b6$sb7$> <ii9apu$ev4$> <ii9gfp$2m0$> <ii9hqs$9q7$>
Reply-To: "Jan Simon" <>
Content-Type: text/plain; charset=UTF-8; format=flowed
Content-Transfer-Encoding: 8bit
X-Trace: 1296598083 27282 (1 Feb 2011 22:08:03 GMT)
NNTP-Posting-Date: Tue, 1 Feb 2011 22:08:03 +0000 (UTC)
X-Newsreader: MATLAB Central Newsreader 869888
Xref: comp.soft-sys.matlab:707535

Dear Bruno,

> Let's assume Jos's bootstrapping does not show an obvious biased, then I believe one still have to show the algorithm works, otherwise to me it is remains a unreliable algorithm.

I'm an experimental physicist. If the cheap bootstrapping reveals a bias, I do not start to think about an expensive proof. I know the fundamental role of prooves, but I keep in mind their limitations, when a prooved algorithm uses e.g. the cheap RNG of Matlab6, real hardware like an Intel processor with a long list of open bugs, and I've heared that even Matlab should have bugs.
Of course you cannot get a reliable program using unreliable algorithms. And you cannot get a reliable program just by implementing reliable algorithms.

And therefore programming combinatorics is fun for *me*.
> You are one of the most rigorous CSSM participant. Almost anything you wrote is sure and exact, and I appreciate a lot this quality. So keep it that way.

Am I rigorous?? After a long hesitating, I cannot decide if this is humorous or a kind of invitation. I'm not used that anybody suggests me how to behave. Anyway, I take it as friendly suggestion with a certain level of humor. I'm going to be very careful if I answer questions or publish functions in the FEX. And I feel free to ask what ever I'm interested in, as long as it concerns Matlab. ;-)

Kind regards, Jan