Re: querying RDF lists [OK?]

On Mon, 2005-10-03 at 19:51 +0100, Miles, AJ (Alistair) wrote:
> Hi all,
> Here is my one comment on the Sparql query language last call [1]:
> I have a specific requirement for being able to represent and access the contents
> of ordered lists within RDF [2].  I would very much like to see at least...
> (a) a standard function (or some other mechanism) for querying whether some node
> is an element of some RDF list (i.e. something like isElementOfList(A, B))

Yes, the WG has considered a number of requests along these lines.
Unfortunately, we have not found any designs that seem
sufficiently mature to standardize at this point.

> ... and preferably also ...
> (b)  a standard function (or some other mechanism) for returning the numerical
> position of a node in an RDF list (i.e. something like listPosition(A, B))
> ... as part of the SPARQL query language WD as soon as possible.  I am concerned
> that the corresponding DAWG issue 'accessingCollections' has been marked as
> 'postponed', and it is not specified at what point this postponed issue
> is to be addressed.  

We don't make any promise that it will ever be addressed. By postponing
the issue, we're just saying that we agree there's an issue, but we
don't agree that there's a requirement to address it in this spec.
It may be addressed by this WG in a future work item, or in a future
WG, or by some other organization, or not at all.

> Speaking more generally from my own implementation experience, this functionality
> is required by almost any semantic web application, and I expect it to
> be a key part of any serious SPARQL implementation.  It would be a pain
> if every SPARQL implementation had a different custom function for doing this.

The WG was polled on whether it seemed essential in the near term,
and many WG members saw "no immediate need for this feature"

Note that some query engines (cwm, euler) are able to infer
list membership triples and if you're using such an engine, you
can query those triples just like any other triples.

a service description approach to accessingCollections

As you seem to be aware, after consideration of requests
like yours and a few designs, on 22 Feb, we
        RESOLVED: to postpone accessingCollections because 
              * our not standardizing it doesn't stop anybody from
              * none of the extant designs seems sufficiently mature
        Clark/UMD, Fukushige/MEI, and 2 others abstaining
I don't see any information in your comment that the WG has
not already considered. I hope you can accept the WG's decision
to postpone this feature until a future work item. If not,
please let us know.

> I'm very happy with everything else :)
> Cheers,
> Alistair.
> [1]
> [2]
> ---
> Alistair Miles
> Research Associate
> CCLRC - Rutherford Appleton Laboratory
> Building R1 Room 1.60
> Fermi Avenue
> Chilton
> Didcot
> Oxfordshire OX11 0QX
> United Kingdom
> Email:
> Tel: +44 (0)1235 445440
Dan Connolly, W3C
D3C2 887B 0F92 6005 C541  0875 0F91 96DE 6E52 C29E

Received on Tuesday, 4 October 2005 22:25:18 UTC