Re: A shot at http://www.w3.org/2000/03/rdf-tracking/Overview.html#rdfms-resource-semantics

Martyn Horner wrote:

> 
> It's clear that the definition of resource in the M&S spec is intended
> to be a `fundamental' definition from which the rest can be built.

Yes, that why its scheduled early.  We need a clear working model for how we
understand resources.  But I'm hoping to avoid a philosophical death spiral
on this one.  We need a model of resources that is sufficient to enable us
to progress.

> It's
> clear also that there's a question whether a `resource' as understood in
> RDF represents the bit-sequence (or the actual atoms) of a referenced
> entity or some sort of pledged token of good intent towards an
> significantly-invariant object.

Its exactly this sort of confusion that we need to clear up.  I don't
believe there is a question here, though there does tend to be some
confusion about this.  For example, RFC 2396 is quite clear that a resource
is NOT the bit sequence (or actual atoms).

> 
> I wonder, given the need to emphasise the legal aspects of RDF
> assertion, whether we don't need to define resources reflexively to
> support this.

Lets try and stay away from the legal stuff if we can.  The law is for
lawyers - we are writing specs - not Acts of Parliament, Bills for Congress,
or whatever the equivalent is in your locale.

> 
> I was playing with a definition along the lines of `a token for an
> object constrained in a set of dimensions expressed by a set of
> properties and values but otherwise capable of redefinition'.

Hmmm.  Might be a good idea to start with whatever definitions already exist.

> There's a
> musing somewhere which talks about successive degress of specificity in
> a web reference. Using `isVersionOf' and `isLanguageSpecificVersionOf'
> allows resources to inherent a set of dimensional constraints from
> another resource or to have them selectively relaxed.

Where?  How authoratative are they?

> 
> It seems to me that, to reflect the spirit of the Semantic Web, we
> should help to define the handles on the system in ways that allow (and
> encourage) fullfillment rather than impose limitations. Defining
> resources by their role in RDF (and not RDF as a process on resources)
> both expresses the spirit and allows a rigorous `external' description
> of `assertion'. It means that `resources' get an RDF-specific
> definition.

Resource is a term widely used in the internet and on the web.  Lets be
careful about confusing everyone by having an incompatible definition.

What I'm hoping for here is something pretty simple that can serve as a
foundation on which we can build.

I recently tried to articulate the working model that I use and it came
out something like this:

  There is a set of resources R.

  Each member of R is identified by a URI by which I mean:

    r1.uri == r2.uri <=> r1 == r2

  i.e. two resources are identical if they have the same URI.
  two resources with different URI's are different resources.

  There is a set E of entities.  Entities are things like web pages,
  numbers and trees in the park.

  There is a mapping D : R -> E.  i.e. there is mapping which maps resources
  to entities.

  There is a notion of equivalence.

  Two resources r1 and r2 can be mapped by D to the same entity.  In that case
  we say they are equivalent.

    equiv(r1,r2) <=> D(r1) = D(r2)

  and finally:

    D((p,s,o)) <=> D(p)(D(s),D(o))

I present this, not because I claim it is correct, it doesn't deal with frag
id's
for example, but to suggest a language in which we can be clearer about
what we mean.

Brian

Received on Tuesday, 8 May 2001 08:43:22 UTC