- From: Richard Cyganiak <richard@cyganiak.de>
- Date: Wed, 25 May 2011 14:59:27 +0100
- To: Peter Frederick Patel-Schneider <pfps@research.bell-labs.com>
- Cc: <jeremy@topquadrant.com>, <public-rdf-wg@w3.org>
On 25 May 2011, at 14:14, Peter Frederick Patel-Schneider wrote:
>> One motivation is to have DATATYPE("foo"@en) return something other than error/undefined in SPARQL.
>
> Umm, isn't that SPARQL business?
SPARQL is not in the business of introducing new datatypes. No existing datatype would be a very compelling answer to that query.
> Also, rdf:PlainLiteral could be used for this already.
That would be possible but unsatisfying. The rdf:PlainLiteral datatype is too unspecific, it's broader than language-tagged literals. And it would be weird if the lexical form of the literal is ill-formed according to the type reported by DATATYPE.
Best,
Richard
Received on Wednesday, 25 May 2011 13:59:58 UTC