Re: [PATCH] Makes sort create random order

[Paul Eggert]

Frederik Eaton <address@hidden> writes:

> 1. It looks like this isn't in 5.2.1 yet, has something been added to
> cvs? Is something planned? Is somebody working on a patch?

Not as far as I know.

Part of the problem -- as you'll see in the thread rooted at
<http://lists.gnu.org/archive/html/bug-coreutils/2004-09/msg00001.html>
-- is that it's a bit tricky to define exactly what a random sort is.

> 2. Was this patch actually tested?

Not by me.

> A better algorithm would be as follows: combine each key you want to
> randomize with some salt (your "random seed"), and compute its hash.

That would be better, but it's hard for me to say whether it
accurately reflected what I wanted.  I would want "sort --random" to
sort according to a random permutation of the correct sort order, with
ties retained.  First, it sounds like you don't want to retain ties.
(I'm curious: why not?)  Second, it's not at all clear to me that the
algorithm you describe will actually result in a random permutation
taken from a uniform distribution of all permutations.

> I think a feature like this would be quite useful to all manner of
> users, from those involved in statistical investigations to those
> who want to randomize their mp3's without installing a separate
> piece of software.

For the latter, almost anything will do; but what sort of statistical
investigations did you have in mind?  That might affect what we mean
by "random sort".  It's quite possible that there should be multiple
flavors of "random sort".

> 3. Is there a way to subscribe to this mailing list?

Yes; please see <http://lists.gnu.org/mailman/listinfo/bug-coreutils>.