Re: RDF 2 Wishlist

On Mon, Nov 2, 2009 at 11:12 AM, Damian Steer <pldms@mac.com> wrote:
>
> On 1 Nov 2009, at 17:51, Sandro Hawke wrote:
>
>> So, what should W3C standardize next in the area of RDF, if anything?
>> OWL 2 added a bunch of stuff to OWL that users wanted and implementors
>> were willing to tackle.  Are there things like that around RDF?
>>
>> My own answer is in a recent blog post:
>>   http://decentralyze.com/2009/10/30/rdf-2-wishlist/
>>
>> What's yours?
>
> I did a quick talk at TPAC last year:
>
> * Deprecate RDF reification. Issue warnings, write document to explain
> problems.
> * Deprecate collections (Alt, Bag, Seq). See above.
> * Serialise all graphs. Let rdf/xml use property URI:
>  <rdf:rel rdf:prop=”http://xmlns.com/foaf/0.1/name”>Damian Steer</rdf:rel>
> * Serialise named graphs (although I'm not super keen in general):
>  * Simple envelope: <document name="foo"
> type="application/turtle">...</document>
>  * Sparql GSPO to dump datasets
>  * etc
> * Deprecate rdf:id. Make about and resource synonyms.
> * Make bnode unlabelled, rather than existentially quantified var.
> * Prefixes: warn if some standard set not 'correct'. Have 'grab all'
> namespace.
> * Lang _and_ type. Reason for exclusivity lost in mists of time.
> * Simple rule language. Not sure RIF has delivered that?
> * Literals as subjects. Not that useful, SPARQL allows it.
> * Bnodes as predicates. See above. Does SPARQL allow it?
> * RDF/XML inverse properties. Make writing more pleasant.
> * Equivalence relations. Seems like every use of sameAs is incorrect.
>

+1 to Damien's answers, and most of everyone's else. By rule language,
he may mean N3. To add in some more clarification to the mix:

1) Instead of reificiaton, look at N3-style and IKRIS-style quoting,
with both a quoting and de-quoting mechanism (ala Lisp). So you can
talk about graphs easily.

2) Agree with Pat/Xie - we need a *single* data-type for plain literal
strings, and all existing other kinds of literals should be coerced
into that. However, would appreciate someone like David making sure
the format has the needed expressive power. Ideally, also then there
should be some standardized way for things like SPARQL to do
type-coercion, so "1" and "1"^xsd:int could be built into language as
an option. Not sure if putting this in RDF Core is the way to go.

3) A JSON serialization, and ideally, a way you can see JSON as RDF,
i.e. as maybe rdf:Lists with a blank subject.

4) Named graphs and collections of named graphs.Need to talk not just
about a single named graph, but collections of graphs residing in a
certain place.

5) Every RDF triple can have a standardized way of tracking provenance
and so establishing trust and changesets..

7) Stop using RDF/XML unless necessary for XML-reasons, instead focus
energies on Turtle-NTriples-N3 sorting out. I think Davifd's already
sorted this out for Turtle/N-triples, just need some W3C stamping and
then an investigation into the stability of N3.

> I also liked Pat's keynote as ISWC. Alas, he didn't have much to say on the
> last issue.
>
> Damian
>

Received on Monday, 2 November 2009 17:04:37 UTC