[
Lists Home |
Date Index |
Thread Index
]
Roy Fielding on WWW-TAG"
"URIs are just names. As such, a URI can identify anything."
That is a leap. URIs are names. URIs can **name** anything.
All that is required is consistent syntax. Try: identification
is a process by which a name is bound to an object. A URI
can name anything and that name can be used in a process
of identification as long as the naming authority behaves
in accordance with expectation (ontological commitment).
If as has been asserted further by Fielding:
"Each URI identifies a resource. The resource identified
is the mapping function, not the result of that mapping, since otherwise
one of the following would have to be true:
1) what the URI identifies is not the resource; or
2) the author has to change the URI every time the result changes.
Since (1) is false by definition and (2) is false by demonstration,
the resource is neither the object on the server (as implemented) nor
the representation (as delivered to the user)."
So:
The resource is the server (practically) and the mapping function (say selector
in the Shannon sense) authoritatively. The resource is the authoritative process
for identification.
Thus, the biases, intentions, philosophies, behaviors, applicable contexts
such as background, social commitment, discussion, negotiation,
and exceptions are proper topics of study of resources. One does
not study the resource by determining what it is but by what it
does. That the essence of ontological commitment as the basis
for trusted communication. It is possibly not a proper field
of study for the TAG.
Again, the URI specification (RFC 2396) should be a syntax-only
specification similar to XML. Names are names and should work
without regard to the binding model as long as it is understood
that to work as network names for addressable objects, they
must be authoritatively bound.
len
|
