Re: Thoughts on the NSA's correction to SHA

[Hal <hfinney@shell.portal.com>]

Bruce Schneier may be correct about NSA's views, but I think the
NSA gives itself too much credit.  There was another very significant
event in the 1970's which IMO played at least as much of a role in
the increased interest in cryptography as the DES.  This is, of course,
the invention of public-key cryptography.

I know that my own interest in crypto can largely be traced to the
Scientific American column by Martin Gardner in which he introduced
the RSA system (along with the famous RSA-129 number which was just
factored).  PK crypto combines simplicity with surprise to produce
results which attracted a lot of attention and interest.  In comparison,
the development of DES was of relatively little interest outside of the
few specialists in the field.  I would suggest that PK crypto did more
to attract attention to cryptography and to lure people to the field than
did DES.

If you look at the papers in the crypto conference proceedings you will
see a number on cryptanalysis of DES and on DES-like systems, especially
in the early days; but there are generally at least as many on PK and
related ideas such as zero-knowledge.  Much of what we think of as
"modern cryptography" owes itself more to the kinds of information
manipulation provided by PK than to the DES, which is often relegated
to the role of a "black box" in a crypto protocol, interchangeable with
IDEA or any other conventional cypher.

It's more defensible to argue that strictly from the NSA's goal of reading
other people's mail, DES was harmful by revealing a general approach for
constructing strong conventional cyphers.  But as far as stimulating the
field of cryptography in general, I think PK has played a more important
role.

Hal Finney
hfinney@shell.portal.com