Re: regarding rdfms-identity-anon-resources

Jonathan Borden wrote:
> Gabe Beged-Dov wrote:

> > There are many ways to signal via a fragment
> > identifier that a particular URIref is a special resource. Using
> > XPointer child sequence is an interesting one but is indirect and
> > requires that these kinds of syntactic shorthands be only used for
> > anon resource labeling.
>         I put forth that all proper labelling of an anonymous resource is
> indirect -- a direct identification is a name. Any resource might be
> anonymous to any particular individual or system, it is all a matter of
> perspective. So yes one may refer to a 'known' resource using an address
> rather than a name. This is no problem.

I am trying to keep my understanding of this issue scoped to the need
to signal that an anonymous resource was described in the RDF/XML
document. This can either be done inband via the URIref that is
generated for the resource or out of band using many possible
book-keeping approaches. You are proposing that the use of a childseq
as the fragment identifier portion of the URI by used to signal this
fact. I would prefer to either use a more explicit signalling
mechanism or do it out-of-band. The out-of-band might either be some
more triples that tracked this metadata or really out of band in some
implementation dependant representation.

> > If you go down this path, I think you might as well use an explicit
> > fragment scheme like x-anon or something else. This also doesn't
> > address round tripping. I.e. how do you regenerate an XML document
> > where the XPointer shorthand (which isn't a valid XML Name it seems)
> > is valid?
>         RDF in general does not address roundtripping of XML documents, to do so
> requires information not stored within the set of triples resulting from
> parsing an XML document. In general complete roundtripping requires an XML
> grove which is also generally a DLG but contains lots of other information
> beyond what is contained in an RDF 'model'.
>         If roundtripping of XML documents is identified as an RDF goal then we can
> have a different discussion.

I have always felt that round-tripping should be not only a goal but a
requirement. It really helps to ground many discussions.
>         The issue of child seqs being invalid XML names is by design! If the
> resource _had_ a valid XML name it wouldn't be anonymous would it?

If anonymity was a first class concept, ala a special variable type,
then you wouldn't have to conflate lack of a name with it being
> -Jonathan



Received on Saturday, 10 March 2001 15:28:40 UTC