On Feb 27, 2006, at 4:23 PM, Hallam-Baker, Phillip wrote:
From: Dave Crocker [mailto:[EMAIL PROTECTED]
... Since the word identity is quite literally central to D*I*X,
the meaning of identity needs to be more than philosophical.
...
I think that what Dave is trying to say here in imprecise fashion is
that we need to have a precise *definition* of what identity means in
the context of DIX and that this must 1) not rely on the outcome of
open
philosophical questions and 2) be understood without direct recourse to
modern continental philosophers.
Yes, precise but more abstract than the syntax of a message so we all
understand the goals.
The sort of problem we do not want to have to deal with is what "is"
is. This is not as easy as it may appear when we attempt to design
technical solutions for abstract things like identity that have such
rich natural semantics.
I believe that a useful working definition is something on the lines
of:
An identity is a set of assertions concerning a particular subject
identifier.
This definition seems to apply to the concept in Dick's ID-2 talk, but
we should be careful. Do we want to say that any set of assertions
concerning an identifier is an identity. This looks like a semantic
trap to me. I suggest that a definition more clearly associated with
what the purpose or use of an identity is might avoid that trap. This
is a trap-rich environment.
This is certainly consistent with Dick's Id2 talk.
The presentation was entertaining. It contained at least one statement
of equivalence that I find unpersuasive from just its assertion. The
equivalence of identity = reputation is a strong and provocative claim.
If the sort of definition of identity on which the WG's effort
(implicitly) rests includes this equivalence, it deserves to be
justified better.
As far as the computer is concerned, it first authenticates 'Dick Hardt
of Sxip', as far as the computer is concerned any machine readable
identifier that binds unambiguously to 'Dick Hardt of Sxip' is
acceptable, for the sake of argument lets use '[EMAIL PROTECTED]'.
Dick's identity is now the set of statements that are bound to
identifiers for Dick:
[EMAIL PROTECTED] is Canadian
[EMAIL PROTECTED] lives in Vancouver
[EMAIL PROTECTED] is Star Alliance member 006342832
Star Alliance member 006342832 is Star Alliance Gold
[EMAIL PROTECTED] likes a bunch of (specified) sports teams
[EMAIL PROTECTED] likes a bunch of (specified) bands
[EMAIL PROTECTED] has blood group Z
[EMAIL PROTECTED] looks like this <dick.jpg>
And so on.
One of the real-world details that is illustrated by this example is
that assertions can be limited. For example, the Star Alliance Gold
might be valid only until January of next year unless Dick flies enough
this year (or has flown way too much already). Since XML is a proposed
format for assertions, it is easy enough to add syntactic elements to
reflect the limitations, but the careful designer will notice the
slippery slope of embedding real-world semantics into the format of
identity assertions.
Careful bounds on the definitions of what we are dealing with are
important here. Relying on the rich set of associations that people
have with an abstract noun like identity will not do.
John
_______________________________________________
dix mailing list
[email protected]
https://www1.ietf.org/mailman/listinfo/dix
