Re: Parsers and rdf:Description

Stefan Kokkelink wrote:

> I agree that the creation of a Bag for each Description
> element isn't necessary in many situations (e.g for
> retrieval) since it represents only the XML structure
> of the RDF document.

I disagree with the last statement. The way I would put it is that the
rdf:Description is the XML serialization for a collection of statements with
the same resource/subject in common AND for the bag of statements that
result from reification of said properties.

Reification is an inherent feature of the model. As such, it has influence
on the XML serialization that we use for RDF, not the other way around.

Now, whether a particular RDF processor exposes reified statements through
its assertion/query API is another story.  I think best practice would be to
do so, but I wouldn't hold it against any processor that chose not to, or
that made it optional.  Some of the fastest XML parsers don't validate--does
that make them wrong/not useful?

If a processor does expose reified statements through the API, I think it is
very useful to allow the client to distinguish reified statements from
"ordinary" statements. Since this is necessary for writing a serialization
anyway, it doesn't cost much in terms of implementation.

Perry

Received on Tuesday, 22 August 2000 13:34:29 UTC