Re: Input sought on datatyping tradeoff

Graham Klyne wrote:

> At 08:53 AM 7/12/02 -0400, Jonathan Borden wrote:
> >The danger in interpreting this idiom in any way other than
> >
> >age = "10"
> >
> >is non-monotonicity. That is in the absence of _some other triples_ i.e.
> >schema, the object of the age predicate is the literal string "10".
> [A late response to your comment;  I've not been following closely, of
> late, and am now catching up email on a long plane flight.]
> A possible interpretation, which doesn't incur non-monotonicity, is that
> Jenny's age is some value that can be represented by the string
> "10".  Which doesn't really tell you anything about Jeny's age in the
> absence of datatype info (hence avoiding problems of non-mon).

Yes. But this seems to say that such literals are "tidy", no?

This is what I mean by suggesting that a literal denotes a _set_ of things,

"10" =>

{ "10", xsd:decimal"10", xsd:binary"10" ... }

and that datatyping info (e.g. a schema) subsets this set.

perhaps this could be called tidy+

in terms of the other proposal:

xsd:decimal"10" =>

_:a1 rdf:type xsd:decimal .
_:a1 rdfdt:value "10" .

or something to that effect...


Received on Sunday, 21 July 2002 10:21:45 UTC