On Mon, 29 Nov 2004 11:08:24 +0900
Wayne Vucenic <nightphotos / gmail.com> wrote:

> Hi James,
> 
> I agree that if we know that 10% of the words are banned, we can do
> better than this.  
> [snip]

I tried to implement an algorithm that is perfect regarding the chunk size given the percentage information. This is described in

http://ruby.brian-schroeder.de/quiz/detect_words/content/content.html
or prettier
http://ruby.brian-schroeder.de/quiz/detect_words/content.ps

The problem is that this algorithm was not significantly better than always dividing into three parts. (Surprisingly it was even slower). Maybe my math is wrong there, if you detect the mind-bug it would be very interesting.

Regards,

Brian


-- 
Brian Schröäer
http://www.brian-schroeder.de/