Re: [xml-dev] ID/IDREF is evil

[Norman Gray <norman@astro.gla.ac.uk>]

Just parenthetically...

On 2014 Feb 20, at 03:55, Kurt Cagle <kurt.cagle@gmail.com> wrote:

> I've been playing in RDF (and RDFA) idioms of late. In general, if I have an outbound resource identification term (an object URI), it is typical in RDF-XML to place it into an attribute, especially if the encoding for the element is not itself RDF. RDF-XML is not my favorite mode of expression - I much prefer working with Turtle - but I'm also trying to be consistent with NIEM conventions. This combination of factors does not necessarily make for friendly XML.  :-)

I think that approximately nobody thinks that the RDF/XML syntax is anything other than sick-making.

There is an argument that its use of XML instance syntax makes it easier to introduce RDF to XML folk; but that argument, while possibly true, runs out of steam very quickly, because the RDF model is quite importantly different from XML's, which is (returning to our starting point) why RDF looks so very bizarre (or indeed perverse) when expressed in this syntax.

The only virtues of the RDF/XML syntax are (i) that it's easier to generate from an XSLT stylesheet, than is Turtle, and (ii) that it would, I suppose, have looked a bit weird if the W3C had standardised RDF back in 1999 referring to a non-XML syntax (I don't know if this was a motivation for RDF/XML in historical fact).

Until recently (last year?) RDF/XML was the sole formally standardised RDF serialisation, which is probably why it lasted somewhat longer than its natural life ('N-triples' was 'standard' but not intended to be used for interchange, and N3/Turtle was a spinoff of a TBL experiment/hobbyhorse/fundamental-initiative).  Turtle is now fully standard, and RDF/XML should now perhaps be relegated to the special cases mentioned above.

All the best,

Norman


-- 
Norman Gray  :  http://nxg.me.uk
SUPA School of Physics and Astronomy, University of Glasgow, UK