Re: RXR, TriplesXML or TriX? Anyone?

On Fri, 14 Jan 2005 11:04:15 -0500
"Brian Manley" <bmanley@granite.com> wrote:

> 
> All,
> 
> Given the current discussion regarding RDF/XML serialization
> alternatives, I was wondering why there doesn't seem to be more
> interest in formats such as RXR, TriplesML or TriX. Each seems
> reasonably sane to me. Is there some flaw in these proposed formats
> that have prevented their adoption?

Hard to say.  I've been too busy to comment in the current thread but
since I edited the RDF/XML refactored spec, created Turtle and RXR,
and written on RDF syntaxes extensively, I think that what's missing
is that nobody found them too compelling (unlike Turtle which has had
several independent implementations).

There is an existing general RDF/XML syntax that is roughly good
enough and more importantly, supported by *all* the RDF tools.  When
people say they want a better syntax, in reality that means they want
a more my-application-focused syntax.  So they usually just make one
(or already have one) and translate to/from RDF triples or to/from
RDF/XML.

In some future standardisation work I could see some just
triples-in-XML like RXR etc. work as useful, but it would be for M2M
as far too verbose for people - all the long URIs.

Dave


Further reading:

A Brief History of RDF Serialization Formats, Oliver M. Bolzer
http://www.fakeroot.net/sw/rdf-formats-20040717/

A retrospective on the development of the RDF/XML Revised Syntax, Dave Beckett
http://www.ilrt.bris.ac.uk/publications/researchreport/rr1017/report_html?ilrtyear=2003