David Alan Black <dblack / candle.superlink.net> writes: > Interesting idea. A first go at this (with, I confess, papered-over > treatment of uppercase letters, i.e., everything is getting downcased > pending some work on that part) is the first threat to the Reign of > Unpack. The following is based on all of /usr/dict/words, using > letter frequencies derived directly from that file: > > user system total real > [first reports removed] > frequency 14.470000 0.000000 14.470000 ( 14.483800) > orig_primes 24.450000 0.010000 24.460000 ( 24.458457) > unpack 15.550000 0.050000 15.600000 ( 15.613195) I think this has been one of the more interesting threads I've seen on a newsgroup. I love the cooperative way you folks are exploring the problem space and refining the solutions. Have you considered writing all this up? I think the idea that (for example) letter frequencies would make a different to an anagram finding program might surprise many readers. We could stick it under the articles on RubyCentral if you wanted to web publish. Thanks Dave