Re: Plain literals and xsd:string

On 2011-04-20, at 18:17, Andy Seaborne wrote:
> 
> On 20/04/11 17:57, Axel Polleres wrote:
>> hmmmmmm, potential effects of RDF decisions on Update or Query at this stage worry me... is there a way we can get around such effects
>> easily without just ignoring the movements in the RDF WG?
> 
> Rather hard to get round, IMO:
> 
> """
> Recommend that systems silently convert xs:string data to plain literals
> """
> 
> and a SPARQL engine isa system.  So we need to consider it even if we find it can be argued for no effect. Ignoring is not really viable.
> 
> I haven't looked at "RDF Entailment Regime" and "RDFS Entailment Regime".
> 
> xsd:string/simple literal is xsd 1a/1b in "7.4 Datatype Entailment Rules" of http://www.w3.org/TR/rdf-mt/.  Don't think it's mentioned anywhere else.
> 
> Doing it right might be modifying rq25.xml#BGPsparql but I really hope we don't have to.  I haven't had space to think about it but it does need checking.  Maybe converting in parsing and/or expression evaluation is enough.

I would expect that's enough, if we can make that change on the basis of a RDF-WG decision. I will change the results of existing SPARQL queries of course.

> On the upside, it makes nice, but tricky-in-the-detail, optimization of filters/patterns with string work.

Yes. I think it will benefit users overall.

- Steve

-- 
Steve Harris, CTO, Garlik Limited
1-3 Halford Road, Richmond, TW10 6AW, UK
+44 20 8439 8203  http://www.garlik.com/
Registered in England and Wales 535 7233 VAT # 849 0517 11
Registered office: Thames House, Portsmouth Road, Esher, Surrey, KT10 9AD

Received on Wednesday, 20 April 2011 17:50:24 UTC