🚀 go-pugleaf

	In the most recent issue of Nature (9 July '92) John Maddox
	discusses a paper from Physical Review Letters (vol. 68, 3805)
	in which a mysterious property of the information content of
	DNA is studied: long-range correlations in nucleotide sequences.
	To put it another way, pieces of information separated by a lot of
	other (related) information affect one another.

	The correlations were found using a random walk model in four
	dimensions and they have a '1/f noise' character.

	This reminded this amateur cryptologist of looking for repeated
	n-graphs to identify the period in a polyalphabetic cipher.  So the
	following questions come to mind:

	Could such an approach be used to identify large numbers used as
	superencipherments?

	Any particular reason why not?

	If you have technical questions, please check the original
	documents, the analysis is certainly is not my field.

	Alec Chambers.

If these are my organization's opinions, I'd be extremely surprised.
 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
<Alec Chambers (jac54@cas.org)  + Americans are a broad-minded people. They'll>
<Associate Editor, CAS          + accept the fact that a person can be an     >
<Phone (614)-447-3600 ext. 3992 + alcoholic, a dope fiend, a wife-beater, and >
<                               + even a newspaperman, but if a man doesn't   >
<Unsolicited Mail:              + drive there's something wrong with him.     >
<P.O.Box.21368                  +                                             >
<Upper Arlington                +  Art Buchwald.                              >
<OHIO 43221-0368                +                                             >
 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^