This has little to do with looking for words by counting every nth character,
as the distribution of first letters of first words of sentences will be much
different than that of words in general. Correct me if I'm wrong, but I think
most words used in the English language start with consonants. The rules of
grammar also limit your word choice (not every word can start a sentence).
This would make it much harder to construct words by sampling letters.
The sample will be much smaller, too, and so it would be harder to get a
statistically valid sample of the population.
You might as well try to create words by taking the first letter of every
word in this posting (to see if Francis Bacon wrote it :-)
--
David Chapman
{known world}!decwrl!vlsisj!fndry!davidc
vlsisj!fndry!dav...@decwrl.dec.com
HOLD IT!
You mean we aren't talking about looking for messages in a particular
*English translation* of the Bible? How sadly this destroys my illusion
of the depth of the meaning of the word: preordained! :-)
(Actually, the algorithm posted by Doug is completely general purpose.
Plugging in your interesting data just makes random word collisions
a foregone conclusion.)
On a lighter note (actually UUCP mail),
sun!uw-beaver!june.cs.washington.edu!gordon (Gordon Davisson) writes:
>The reason you and Doug got different numbers is that you're looking
>for a probable match and he settled for P = 0.1
My apologies for both misunderstandings.
--
Istvan Mohos
{ihnp4,decvax,allegra}!philabs!hhb!istvan
HHB Systems 1000 Wyckoff Ave. Mahwah NJ 07430 201-848-8000
====================================================================
The classic chemical engineering text, "Transport Phenomena" by Bird,
Lightfoot, & Stewart contains numerous instances of (intentional!)
hidden acrostics and other word play.
By the way, I did that since the crude assumptions I was making
would break down pretty badly if the probability were high. An
accurate general analysis could be made but I didn't want to
work it out when a quick estimate was possible..