Re: Loops (RE: CPIM changes)

Derek,


well, I was trying to distinguish between "can" and "have", ie, theory vs.
practise.  given the problematic history of fielded security systems, 'have'
is more persuasive, by quite a lot.

Theoretically, the compress/decompress does qualify as a successful example
of alternative representations, but I was hoping there was an example that
really did translate underlying content syntax.

It occurs to me that an example of the problem in transfering this bit of
history to the current discussion is that each end needs to work on the
theoretical canonical form, no matter what local representation is.  They
might be expected to do that correctly for compressed data, but not for
arbitrary syntactic translation.  So we are back to a system that absolutely
requires end-to-end knowledge of CPIM.

d/

Saturday, November 16, 2002, 12:35:58 PM, you wrote:
Derek> Dave Crocker <[email protected]> writes:
>> interaction between systems that do different encodings of the canonical
>> form have actually been fielded?  that would be seriously cool!

Derek> I think we're talking past each other about "canonical form".  In the
Derek> PGP case, there is a well-defined "this is how you take a text message
Derek> and transform it before taking the hash for a signature".  What you
Derek> get as output from PGP at that point is a standalone signature.  You
Derek> can then send the text message by any lossless method along with the
Derek> signature, and the signature is verified on the other end.

Derek> For example, I can take a text message, text-sign it, then compress
Derek> the message, uuencode it, and email you the uuencoded, compressed
Derek> message and the PGP signature in two MIME parts.  You can then verify
Derek> the signature after you uudecode and uncompress the message.




d/
-- 
 Dave Crocker  <mailto:[email protected]>
 TribalWise <http://www.tribalwise.com>
 t +1.408.246.8253; f +1.408.850.1850




  [reminder: [email protected] for non-technical discussions, please]