1993-01-27 - Computerized OTP (was 5th AMENDMENT & DECRYPTION)

Header Data

From: tribble@xanadu.com (E. Dean Tribble)
To: uunet!GS80.SP.CS.CMU.EDU!Marc.Ringuette@uunet.UU.NET
Message Hash: 9c2ea99ff03d19250085e7c5dfbf07406466525a65519402db81fe0e95ccee41
Message ID: <9301270816.AA08283@xanadu.xanadu.com>
Reply To: <9301270519.AA17681@toad.com>
UTC Datetime: 1993-01-27 08:50:22 UTC
Raw Date: Wed, 27 Jan 93 00:50:22 PST

Raw message

From: tribble@xanadu.com (E. Dean Tribble)
Date: Wed, 27 Jan 93 00:50:22 PST
To: uunet!GS80.SP.CS.CMU.EDU!Marc.Ringuette@uunet.UU.NET
Subject: Computerized OTP (was 5th AMENDMENT & DECRYPTION)
In-Reply-To: <9301270519.AA17681@toad.com>
Message-ID: <9301270816.AA08283@xanadu.xanadu.com>
MIME-Version: 1.0
Content-Type: text/plain


	 "All the properties you want?"  What you want is random, and nothing else!

	 all previous bits seen, the probability that the next bit seen will be
	 zero or one is exactly 0.5.  

Note that in practice, the length of a string of 1's or 0' is
irrelevant:  The chance of a string of length N being all the same is
O(2^N), so becomes unlikely for reasonably short strings of bits (1 in
1024 for 10 bits), and virtually impossible for interesting sizes of N
(1 in 4 billion for 32 bits).  This doesn't even strike as being worht
the effort of figuring out how badly the OTP is compromised by
shortening such runs.

Remember how badly our intuitions are on things like security.
Believe the numbers, not your gut feel.

dean





Thread