Re: thinking about the formal model for RDF

Hi William,

On Wed, 18 Jul 2001, William Loughborough wrote:

> At 01:51 PM 7/18/01 -0400, Dan Brickley wrote:
> "the RDF name for a resource is its URI", not "we can't mention resources
> without providing their URI"
>
> What does "mention" mean here? I've been operating under the (naive?)
> delusion that to mention something *is* to name it. Just what form exactly
> does this "mentioning" take?


An example snippet, lots of mentioning, not much naming:

<Company>
  <owns>
    <Company>
      <owns>
        <Company>
          <employee>
            <Person>
              <hasSister>
                <Person>
                  <shareOwner>
                    <Company>
                      <!-- etc... -->

This "striped nesting" of elements alternates between describing some
resources (some companies and some people) and some relationships that
hold between them. It shows how RDF lets you mention a whole lot of
resources, only some of which you  might be bothered (or knowledgeable
enough) to describe in any detail. Providing URIs is just another part of
the descriptive process; it's a particularly crisp and useful way of
describing (and identifying) things. Anonymously mentioned resources in
RDF are sometimes described in enough detail to uniquely pick them out.
More often than not, though, the description is not in itself enough to
tell which thing is being described. For example, it would be consistent
with the example above for the employees sister to be a shareholder in
the parent company, but we can't tell from the data given which company
she holds shares in.

Does that help?

Dan

Received on Wednesday, 18 July 2001 15:15:06 UTC