crypt(1) -- how secure, how breakable?

[Henry Spencer]

A friend has asked a couple of questions about the security of crypt(1)
that I wasn't able to answer.  I know that crypt(1) has been broken,
although the software for it is not generally available and is not
trivial to use, but he's interested in some specific information for
a customer.  In particular...

	(1) Are there any papers in the open literature discussing
		the security/breakability of rotor machines?  I had a
		vague memory that Cryptologia has published something
		like this once, but don't have specifics.

	(2) Apparently there is some sort of rating scheme for encryption
		systems, based on estimated breaking time on a Cray or
		something like that.  Or so he thought.  Has anyone ever
		rated the crypt(1) algorithm on such a scale?

Please reply by mail; I will summarize results for the net.
-- 
				Henry Spencer @ U of Toronto Zoology
				{allegra,ihnp4,linus,decvax}!utzoo!henry