Re: varieties of datatyped tagged literals

On Sep 8, 2011, at 9:50 PM, Richard Cyganiak wrote:

> On 9 Sep 2011, at 02:51, Pat Hayes wrote:
>>> The *literal* has a lexical form.
>>> The *datatype's* lexical space is empty.
>>> Consider "hello"^^xsd:integer. That's a typed literal and it has a lexical form "hello". That lexical form is not in the lexical space of the literal's datatype.
>> So the literal is ill-typed. One cannot have a well-typed literal with a lexical form when the datatype has an empty lexical space. Yet there are, apparently, well-typed literals with this type, and (you tell me) they have a lexical form. Which ought to be impossible.
> The notion of ill-typedness does only apply to literals whose value is determined using an L2V mapping.

No, applies to any literal with a datatype. And in any case, at present, ALL typed literal values are determined by an L2V mapping: that is part of the very definition of a datatype. So this strange device which you are wanting us to use would be just as much an exception, requiring a special explanation and exceptional handling in the semantics, as any of the other options. And IMO it is so unintuitive as to be more confusing than helpful. It would be simpler and easier to understand to simply specify by fiat that plain and tagged literals have certain rdf datatypes, ignoring the L2V material, and then write these as exceptions into the 'datatype' section of the semantics.  


IHMC                                     (850)434 8903 or (650)494 3973   
40 South Alcaniz St.           (850)202 4416   office
Pensacola                            (850)202 4440   fax
FL 32502                              (850)291 0667   mobile

Received on Friday, 9 September 2011 15:28:40 UTC