From: Bill Stewart <stewarts@ix.netcom.com>
Date: Sun, 14 Apr 1996 10:20:02 +0800
To: "Declan B. McCullagh" <declan+@cmu.edu>
Subject: Re: Enforcing the CDA improperly may pervert Internet architecture
Message-ID: <199604100704.AAA21801@toad.com>
MIME-Version: 1.0
Content-Type: text/plain


One issue with the "We could require that everybody label their packets"
proposal is that a large part of the world isn't under FCC or even US control;
why should some foreigner label their packets based on the tastes of US censors?

Another problem is that adequate labelling requires public-key cryptography,
and the US bans export of crypto; this means you can't use the best technical
standards for domestic use and export, and means you can't mass-market
exportable rating software.  We EEEVILLL Net Users would certainly
be _happy_ if the FCC or Congress talked the Administration into legalizing
the use of decent authentication technology.  (They could argue that the ITAR
permits authentication-only technology, but there are clear technical 
advantages to RSA vs. DSS, and DSS has the subliminal-key options that
mean you can use it for non-authentication encryption as well as signatures
anyway.
There's also the problem that both are patented, though the patents behind
DSS are weak and run out in a year or two.)

Another problem is that this proposal would require multiple authentication
headers per IP packet - not only is it wasteful, but is it even supported?
I suppose there's some tunneling approach possible, but it'd be really awkward
and non-portable.

ObExon:  Does the Administration propose to label any on-line copies of 
the Federal Register?  There's often material in there, such as the recent
Congressional debates on partial-birth abortion, that are clearly in violation
of the CDA if posted to the nets, so they would have to be labeled.
#					Thanks;  Bill
# Bill Stewart, stewarts@ix.netcom.com, +1-415-442-2215