That was my point exactly

From: Paul Tomkins (
Date: 10/06/04

Date: Thu, 7 Oct 2004 07:05:28 +1000

>> Assume that both the message that you want to encrypt and the key you are
>> using are both English sentences or English words.
> Well, that's all well and good, but that violates the requirement that the
> key is truly random. Moreover, it says that you know something about the
> plaintext that is not evident from the ciphertext.

My point was that you can't use English words of sentences as keys in a
OTP-type cipher because that violates the requirement that the key is truly
random. (I am in perfect agreement with you.) I then preceded to show how to
crack a OTP-type cipher that did use English words or sentences as the key
and how vulnerable such encryption is (hence my use of the phrase "Assume
that both the message and the key are English words or sentences.").

I was not saying that in general a OTP-type cipher would use an English word
or sentence as the key, that the cracker would therefore be able to use that
assumption, or that OTP-type ciphers are breakable because both the
ciphertext and key are English words or sentences.

A OTP-type cipher **must** use a random key in order to be secure. There was
a lot of talk of cats and dogs in posts so, even though a number of people
pointed this out in their posts, I was concerned that Gustavo may have held
and continued to hold the opinion that any key was acceptable even if it was