From: Hal Finney <hal@rain.org>
Date: Mon, 30 Dec 1996 09:59:31 -0800 (PST)
To: cypherpunks@toad.com
Subject: Re: New crypto regulations
Message-ID: <199612301757.JAA00424@crypt.hfinney.com>
MIME-Version: 1.0
Content-Type: text/plain


From: Mike McNally <m5@tivoli.com>
> solman@MIT.EDU wrote:
> > The government's claim is that in the interests of national security,
> > export of cryptography must be prevented. By limiting the policy's
> > applicability to media which are in, or can easily be converted to,
> > electronic form ...
>
> Does anybody seriously believe that nbody writing these policies has
> an understanding of OCR software?  An on-line form of code printed
> in a book is just a quick trip to a scanner away.  They know that.

The regs, as Lucky pointed out, do hint at restrictions on OCR fonts in
the future.  However this is obviously doomed since as OCR technology
advances the distinction between OCR and non-OCR fonts will vanish.
I imagine that a special purpose character recognition engine could be
built to work on any known, monospaced font, as is typically used for
source code.

In this light, the explicit exemption for printed materials is really
quite welcome.  It has never been 100% clear that a book of source code
is exportable.  Yes, we've had some favorable court cases recently but
none of these have been fully resolved.  Rumors were posted here that
the NSA came very close to trying to stop the export of the original PGP
source code book from MIT Press (and supposedly arranged for MIT to be
punished later for its audacity).

Having all sides agree that crypto source code can be exported in printed
form is an important step in the right direction.  We can still contest
the issue of restrictions on machine readable exports.  In an era where
electronic publishing is becoming as important as paper publishing for
expressing ideas, we can continue to push to extend the exemption to
machine-readable images of the pages of the book, and later to actual
source files.

Hal