[Cryptography] mathematical constants

[Peter Gutmann]

Jerry Leichter <leichter at lrw.com> writes:

>What would work would be pre-commitment to values.

How many attacks have there been due to published booby-trapped values in
crypto algorithms?  (EC-DRBG doens't count because the private-key values were
never made public, I mean published definitely non-NUMS values).

Now, how many attacks have there been due to buffer overflows, XSS, SQLI,
stack-smashing, ... ?

On the way back from a recent security meeting a friend of mine made the
comment that worrying about side-channel attacks (which had come up there) was
like frantically phoning around plumbers to fix a loose tap in a house that's
on fire.

Peter.