"hitesh.jasani / gmail.com" <hitesh.jasani / gmail.com> writes:

> Actually if you flip it around as 'leagues & words.split' it turns out
> to have some significant performance advantages in many cases.  See
> http://www.jasani.org/articles/2006/02/28/adding-the-science-back-to-computer-science
> for more details.

You'd better cache those Regexps.
Also, test with longer "words"---they are more likely to grow than the
number of leagues.

> - Hitesh
-- 
Christian Neukirchen  <chneukirchen / gmail.com>  http://chneukirchen.org