Why doesn't Sun release the crypto module of the OpenSPARC? Crypto export restrictions

[John Gilmore]

> I would expect hardware designs to be treated more like hardware than 
> software.

A hardware "design" is not hardware.  Only a naive parsing of the
words would treat it so.  A software design is not treated like
software; you are free to write about how ATM machine crypto is
designed, even if you can't export ATM machine crypto software without
a license (because it's proprietary and not mass-market).

A hardware design is a lot like software.  It's human written and
human readable, it's trivial to reproduce, it's compiled automatically
into something that can execute, and if you write it into hardware,
then it does something.

The court case that EFF won against the export controls was won on
those grounds: the government can't suppress the publication of
human-written and human-readable text, on the grounds that somebody
somewhere might put it into a machine that does things the government
doesn't like.

Sun may be chicken on the point, and the government did a sneaky trick
to technically avoid having a Ninth Circuit precedent set on the
topic, but a similar precedent was set by Peter Junger's case in
another circuit.  I think Sun would be well within its rights to ship
VHDL or Verilog source code that implements crypto under an open
source license.  And I'd be happy to point them at good lawyers who'd
be happy to be paid to render a more definitive opinion.

	John Gilmore


---------------------------------------------------------------------
The Cryptography Mailing List
Unsubscribe by sending "unsubscribe cryptography" to majordomo at metzdowd.com