Re: Comments on SPARQL draft

On 04/11/2004, at 2:10, Seaborne, Andy wrote:

> On the other hand, there is much RDF data that does not use typing yet.

I think the best solution here would be to provide a function that can 
map between lexical forms and datatyped values.  This allows people to 
query untyped RDF and designate the type where the author of the RDF 
hasn't.

> I propose that SPARQL does not require processing plain literals as
> numbers if they just happen to look like numbers.  Implementations 
> would
> be free to provide this if they choose to but it is outside the rec
> (i.e. SPARQL does not forbid it either).

I'll be unhappy if we decide that an operator is required by the SPARQL 
spec, but that its correct behavior then wasn't precisely defined.  
Implementations should be free to add operations that do their own 
weird comparisons via the extension mechanism, but we shouldn't be 
leaving them any doubt when implementing operators that get hardcoded 
into the language.

Received on Friday, 5 November 2004 01:18:24 UTC