On Thu, 19 Mar 1998, Anonymous wrote:
I think the answer must depend on why we want symmetric algorithm 0.
At some level it might be a security risk to have it there: if an
attacker can get the naive user's configuration file to specify that
algorithm as the default encryption algorithm and the CFB/IV system
were used as tzeruch and Section 5.7 suggest, there would be no easy
way for the user casually looking at an encrypted file to know that
it was unprotected. This would also be true if an Algorithm 0 packet
were left clear (i.e. without random IV and CFB) but compressed.
Is Algorithm 0 specified for debugging? If so, do we need to say it
MUST NOT be enabled outside of debugging environments?
I ended up implementing it without the CFB (though with the 10 byte reset
IV), so I could use less or od -a to see the not-really-encrypted text.
The same question still holds for Algorithm 5, ROT-N: should we
make that "MUST NOT be enabled outside of debugging environments"?
I did this without CFB, So every space maps to the same character, etc.
So the structure of what you encrypt is visible. I also added a cfb
ROT-8-n variant which does mess things up a bit more.
Beyond debugging, it creates a diagnostic mode, but you are right that
these should be disabled in such a way that it is very hard for a user to
enable them.
--- reply to tzeruch - at - ceddec - dot - com ---