Re: 'in the result of'

On Tue, May 26, 2009 at 12:08 PM, Seaborne, Andy <andy.seaborne@hp.com> wrote:
>
>
>> -----Original Message-----
>> From: Sandro Hawke [mailto:sandro@w3.org]
>> Sent: 26 May 2009 16:51
>> To: Seaborne, Andy
>> Cc: public-rdf-text@w3.org
>> Subject: 'in the result of'
>>
>>
>> You wrote:
>>
>> > Maybe (intro):
>> >
>> > """ ...typed rdf:text literals MUST NOT occur in
>> > published RDF content or in the results of SPARQL basic graph pattern
>> > matching [SPARQL] using extended SPARQL Basic Graph Matching"""
>> >
>> > *in the result of* is the key here.
>>
>> which doesn't seem right to me.  Don't you instead want to rule out the
>> SPARQL queries being against a graph using rdf:PlainLiteral?
>>
>> Are you really okay with this:
>>
>> graph:
>>      <a> <b> "Hello@en"^^rdf:PlainLiteral.
>> query:
>>      SELECT ?x WHERE { ?x <b> "Hello@en"^^rdf:PlainLiteral }
>> }
>
> Fairly OK - I'm not so worried about this situation - others might be.
>
> It would take an app writer who has knowledge of rdf:PlainLiteral to do it hence OK to me.  I do care about data exchange and don't like the two representations of the same thing that existing systems will see as different.

I'd be concerned, I guess. Exchanging queries happens too. For a
system that knows how to do d-entailment writing the query as

SELECT ?x WHERE { ?x <b> "Hello@en" }

will pose no problem. However if you are on the receiving end of a
query that has the rdf:PlainLiteral in it you are guaranteed that it
won't work if your system doesn't do the right sort of entailment.

-Alan

>
> The case that need addressing is in what happens to results of BGP matching, you plug a SPARQL/OWL BGP extension into an existing engine and ^^rdf:PlainLiteral come out (in SPARQL Query Results in XML) and queries that did work may not.  This, to my reading, is not covered by "published RDF content" (it's not 'publishing' to me) so I suggest explicit text to deal with it.
>
>>
>> I thought this was part of what you were trying to rule out.
>>
>> It seems to me the text you've proposed allows that, but happens to
>> (oddly) rule out stuff like this:
>>
>> graph:
>>     <a> <b> "Hello"@en.
>> query:
>>     CONSTRUCT { ?x ?y "Hello@en"^^rdf:PlainLiteral }
>>     WHERE     { ?x ?y "Hello"@en }
>
> I believe that is already covered by the "published RDF content" text.  It is "MUST NOT".
>
>>
>> which doesnt seem like something we need to worry about.
>
> We do.  It creates a graph with "Hello@en"^^rdf:PlainLiteral which wil not be understood as "Hello"@en by deployed systems.
>
>>
>>      -- Sandro
>
>        Andy
>
>

Received on Tuesday, 26 May 2009 17:37:00 UTC