RE: comments on syntax wd: bug in graph seriali[zs]ation algo

On Fri, 26 Apr 2002, Jeremy Carroll wrote:

> DanBri:
> > I propose section 6 be dropped for now, until this is fixed.
>
> Opposed.

Ah, we disagree.

> There are practical problems in RDF serialization.

Agreed

> At some point the specs need to acknowledge this, and indicate to
> implementors what they should do.

Offer guidance, sure. Provide pseudocode -- perhaps in a Note.

The current text provides an algorithm for taking an RDF graph and turning
it into an RDF/XML document whose meaning is different from the graph's.
Our decisions on bNodes make this clear.

> In RDF2 I would hope to fix this, e.g. allow the "_:nnn" bnode syntax as
> matching uriref production in the grammar.
> That was out of charter.

Yup.
I'd do it a bit differently: unique bnode labels could be (just about...)
done as unambiguous properties of a resource. (as could URI names, imho,
but definitely RDF 2 territory there).

This doesn't help folk tying to serialize the 1.0 spec graphs, and having
guidance in the doc (or a doc) is useful. I want to write a serializer in
Ruby, for eg., and guidance much appreciated.


> The charter allows us to clarify M&S.
> The clarification is "this is screwed up".

Having a serialization algorithm that directly contravenes one of our most
painfully drawn out decisions also feels a bit wrong. I'd rather the spec
stayed silent or we introduced a better version of the id/idref linking
syntax (which'd take time) than include the current spec.


> I think a minor change highlighting that the meaning of the graph has
> changed in such a serialization may improve the document.

If the meaning changes, it's not a serialization so much as a
transformation...

Dan
>
> Jeremy
>
>
>
> > -----Original Message-----
> > From: w3c-rdfcore-wg-request@w3.org
> > [mailto:w3c-rdfcore-wg-request@w3.org]On Behalf Of Dan Brickley
> > Sent: 26 April 2002 12:14
> > To: w3c-rdfcore-wg@w3.org
> > Subject: comments on syntax wd: bug in graph seriali[zs]ation algo
> >
> >
> >
> > http://ilrt.org/discovery/2001/07/rdf-syntax-grammar/#section-Serialising
> > [[
> > All blank nodes are assigned arbitrary URIs.
> > ]]
> >
> > This is inconsitent with our policy w.r.t. blank nodes.
> >
> > I propose section 6 be dropped for now, until this is fixed.
> >
> > Dan
> >
> >
>

Received on Friday, 26 April 2002 11:36:23 UTC