Lists Home |
Date Index |
> So in your RDF blob (the gizmo containing the triples) you would
> each triple as a representation of that resource. Good idea. It sounds
Well, that's the reciprocal of what I said.
I said that "any representation is a statement about a resource". I did
not say that "any statement about a resource is a representation".
I would be uncomfortable going that far. Representations are really a
subset of the kinds of statements which can be made about a resource.
> then you need a filter to make it an RDF triple since it doesn't come
True. I don't even know how valuable that would be anyway. I'm
certainly not advocating that representations be stored as RDF.
For example, a hyperlink inside a page is an assertion about the page as
well; so if http://www.ibm.com homepage has <a
href="http://www.microsoft.com">MSFT</a>, you could conceptualize this
(or any number of variations; some which could contain the label, etc.)
But again, all I am saying is that "any hyperlink is a statement about a
resource". Conceptually, this is true, but I wouldn't advocate getting
rid of hyperlinks and storing them as RDF.
> Anyhow, if we refer to RDF statements as representation I am for it,
Again, I wouldn't go that far. A representation served through HTTP is
a statement about a resource, but it's a very specific type of statement
(one about the way the resource is intended to be depicted in a mime
viewer) and is rarely if ever made with RDF. A representation is like
the (o) of an (s, p, o) triple, and it's rarely associated with the (s)
through RDF. Instead the representation and resource are tied together
through HTTP protocol. I like that foaf calls things "depiction", while
HTTP calls them "representation". Although they are conceptually doing
the same thing, it's smart to keep "representation" in the realm of