SV: Copyrighting schemas, Hailstorm

[Dimitris Dimitriadis <dimitris.dimitriadis@improve.se>]

comments inlined

If you don't care reading the whole mail, the answer to the question of
whether implementors can claim ownership of means of identification is no,
and that it should be public.

However, I do have some questions, so I'd really appreciate it if you looked
through it.

-----Ursprungligt meddelande-----
Från: Bullard, Claude L (Len) [mailto:clbullar@ingr.com]
Skickat: den 1 juni 2001 21:27
Till: Dimitris Dimitriadis
Kopia: XML DEV
Ämne: RE: Copyrighting schemas, Hailstorm


Ok, but a term for which we have no useful 
definition does not provide much that is useful.  
This dips into the philosophy of self and that is irrelevant. 

[dd] Which it the term for which you have no useful definition? And why
would philosophy of the self be irrelevant? Are you absolutely convinced
that our discussion has absolutely nothing to do with concepts of the self,
of which you are presumably not particularly fond? I take it that it is the
very concepts in, among other disciplines, philosphy of self, that by far
precede any technical and implementation specific discussion.

Identification or assertion of identity is an 
act and auditing the act is doable.  One act 
asserts identity and another asserts the 
correctness of the assertion (authentication). 

In a wide open system, ensuring 
the actor owns the asserted property is very 
difficult.  That is why the most highly secured 
systems are in vaults and have no external 
connections.  Protection is commensurate with 
risk.  As individuals risk more, they want and 
need more protection.  Hailstorm is MS trying 
to both enable more services across more platforms 
while enabling more protection.   The question 
is does it work?  Then who works it? 

[dd] The question is not if it works. Everything works if properly tweaked.
The question is if we want it to work. And in case we do, the answer to your
last question is "whoever we allow to work it". Simple as that. And that is
where I thought the argumentation took off.  Besides, I gave you a perfectly
good example of where the kind of information that resides in a computer in
a vault ended up on my computer on my desktop in an office (the insurance
example).

But the original question is if the means is 
system wide, can they assert ownership of means.

[dd] If I have a particular set of identification means (password, voice,
retinal scan, fingerprint, what have you) and rest assured that that's
enough and they can be forged and used by others, we end up in the pig
loving donkey case (only difference being that I have less money and more
bills, possibly even a secret lover I didn't know of until then). If, on the
other hand, we can come up with alternative means that cannot be forged, we
can rest assured that nothing bad will happen. 

[dd] I don't know if those alternative means can be made secure enough to
work in an environment of distributed, fully readable, text based,
descriptions of services. This non-forgeable set of means would have to be
public, yes, in order for the publically authorized party to play the role
of authenticator. So they (the other guys) could not claim ownership, no.

[dd] How this would work is a totally different question. Except in the
case, of course, where it is made manually by a bunch of employees in a
vault somewhere running between computers that are online and computers that
are not and press buttons according to whatever info they have on their off
line machines to validate incoming transaction requests. But that doesn't
sound likely.

[dd] So, the question really becomes: Can we think up a series of digital
authentication means that are sound enough to use as we would use our ID or
passport today? If so, would a company's ownership (or use) of those be
hazardous as to the integrity of the person identified using them? In short,
SHOULD they have ownership? And that is the question to which my answer is
no. It is up to someone else to discuss the technical details of
implementing it.

/Dimitris