Re: safety of blindly truncating hashes (Re: TEA)

Alex Alten (Andrade@netcom.com)
Thu, 02 Jul 1998 23:03:08 -0700

• Next message: Adam Back: "Re: safety of blindly truncating hashes (Re: TEA)"
• Previous message: Matt Blaze: "more cryptanalysis of Skipjack variants by Biham et al"

At 10:29 PM 7/1/98 +0100, Adam Back wrote:
>
>Alex Alten writes:
>> You can't just truncate blindly. You need to understand how the
>> hash is internally constructed and operates. On paper you may be
>> improving the strength in one area, even in the limited case you
>> mention above, however in fact you may have reduced it too much in
>> another.
>
>Possibly for broken proprietary hashes the collision resistance and/or
>diffusion is so broken that truncation of removal of certain bits
>would be worse than removal of other bits. Probably this is what you
>were commenting on originally. ("you can't truncate blindly").
>

This is what I meant. To improve performance a designer may chose to
have less than even diffusion as long as he could guarantee a minimum
set of bits would change. If you happen to truncate in these bits
the hash may be useless over certain ranges of text input.

- Alex

--
Alex Alten
Andrade@Netcom.Com
P.O. Box 11406
Pleasanton, CA  94588  USA
(510) 417-0159

• Next message: Adam Back: "Re: safety of blindly truncating hashes (Re: TEA)"
• Previous message: Matt Blaze: "more cryptanalysis of Skipjack variants by Biham et al"