From: Chris Pine <nemo@...>
Date: 2003-04-21T22:57:36+09:00
Subject: Re: Biased weighted random?

----- Original Message -----
From: "Mauricio Fern�ndez" <batsman.geo@yahoo.com>

I see that
   max(Ni) > sum(Ni)/2 => p(Vi) > 0.5 => would have to repeat
however I don't know why you have '+1' (unless you meant '<') :-?
----------------------------

I think he's solving a different problem; he's not using weights, but
numbers of occurances.

So, for example, if he has {'a' => 1, 'b' => 2}, he will give:

  bab

but that's the end of it!  So the `+1' is because it can be at both ends.

Interesting, but it is a different problem:  we weren't looking for a
shuffle (at least, I didn't think we were; perhaps I'm wrong!), we wanted an
arbitrarily long sequence.

Chris