'all resources' is not a set

Another little suggestion for theoretical underpinnings
of URIs: I suggest abandoning any formalism where
'resources' form a set. 

Defining resources as a class rather than as
a set has several advantages. With a class, there
is no need for a well defined equality. The
"same" relation does not have to be well defined
for all pairs.


You get to eschew trying to have a well-defined notion
of equality amongst resources. When asked 'can
two different resources have the same URI', you can
counter that the question is ill-formed, because there
is no way, in general, to tell if two resources are
the 'same' or 'different'. Is the referent of
'the morning star' and of 'the evening star' the same?
Not defined.

Now, there is, of course, the set of all URIs, and
a (more or less) well defined equality relationship
amongst URIs.

There are still some axioms that can be stated without
having equality. One axiom you may want is to assume
that, whatever the 'meaning' is for a URI, it doesn't
depend on time, the observer, or the state of the
observer.

But avoid treating resources as if they formed
a set.

Larry

Received on Thursday, 25 September 2003 17:17:44 UTC