Re: Another history question was Re: Tunny and SIGABA
From: Tony T. Warnock (u091889_at_cic-mail.lanl.gov)
Date: 06/09/03
- Next message: clem: "Re: The TDCAL License"
- Previous message: Brian Gladman: "Re: Avoiding C++ Templates In Cipher Implementation"
- In reply to: Kent Betts: "Re: Another history question was Re: Tunny and SIGABA"
- Next in thread: Joe Peschel: "Re: Another history question was Re: Tunny and SIGABA"
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ] [ attachment ]
Date: Mon, 09 Jun 2003 08:04:48 -0600
Kent Betts wrote:
...
> For a long time I was thinking the way the Enigma never allowed a letter to code
> as the same letter was a design flaw, since it was a way into the machine. It
> finally dawned on me that this aspect of the Enigma was probably done
> intentionally to provide convenience to the operator. ( Having the coded letter
> showing up as the plaintext letter would have been annoying.)
However, this is a very big security hole. If one had a message in depth, then one
gets information such as: the first letter is neither T nor A nor G, and similarly
for other letters. A maximum likelihood analysis can be constructed given the
constraints and this makes things easier. For example in English, the most likely
3-letter word is THE but if the first letter is know not to be a T or an A, then
both THE and AND and ARE are excluded.
- Next message: clem: "Re: The TDCAL License"
- Previous message: Brian Gladman: "Re: Avoiding C++ Templates In Cipher Implementation"
- In reply to: Kent Betts: "Re: Another history question was Re: Tunny and SIGABA"
- Next in thread: Joe Peschel: "Re: Another history question was Re: Tunny and SIGABA"
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ] [ attachment ]