RE: Change in definition of RDF literals

...
> PROPOSE
>  ...
>   rdf:XMLLiteral datatype is changed to have the identity as 
> its lexical
> value mapping (no wrapping), with consequential change to the 
> value space of
> rdf:XMLLiteral.

Are you then presuming that the lexical form in the graph is
canonicalized?

Would it not be better if the canonical form was the value, and
the L2V mapping was the canonicalization? Since equality is
defined in terms of canonicalized forms.

This seems more in line with the nature of datatypes, where
there can be more than one lexical form that maps to the same
value -- thus one can express an XMLLiteral in several ways
all of which canonicalize to the same form.

I think it is wise to have the abstract graph retain the
lexical form as expressed in the RDF/XML, just as is done
for all typed literals, and treating canonicalization as
the L2V mapping for XMLLiteral would accomplish that nicely.


Patrick

Received on Wednesday, 14 May 2003 01:46:24 UTC