Re: ISSUE-19: Should TURTLE allow triples like "[ :p 123 ]." as SPARQL does ?

From: Richard Cyganiak <richard@cyganiak.de>
Subject: Re: ISSUE-19: Should TURTLE allow triples like "[ :p 123 ]." as SPARQL does ?
Date: Thu, 31 Mar 2011 23:01:11 -0500

> On 31 Mar 2011, at 17:57, Peter Frederick Patel-Schneider wrote:
>> I don't see any inconsistency.  You appear to want to move something
>> that is a node and put it in a place where a triple is expected?
> 
> [ <c> <d> ] is not a node. It is three nodes.

[ <c> <d> ] is indeed a single node, just like 5+7 is *one* integer, not
three.  I don't expect to be able to use 5+7 as the single argument of
function that has arity two, or three.  I similarly don't expect to be
able to use the blank node resulting from [ <c> <d> ] in a place that
expects a triple.

>> Why should that work?
> 
> Why shouldn't it? It works in N3 and SPARQL.

I rest my case.   [Note, no smiley here, particularly for N3.]

As far as I can see neither
http://www.w3.org/TR/rdf-sparql-query/
nor
http://www.w3.org/TR/sparql11-query/
allow this construction as a triple anyway.  So I don't see how it works
in SPARQL.

>> If this change is made, then constructions should also be so promotable.
> 
> What is a construction?

Sorry, I meant to say "collection", i.e., a list.

> Richard

peter

Received on Friday, 1 April 2011 14:48:39 UTC