>Sorry if I've missed something in my struggle to catch up, but in what
>sense is xsd:token the builtin token type? I thought we were
>datatype-vocabulary neutral, allowing a binding to xsd if desired?
I have proposed that we have two builtin datatypes, token and string, with the same semantics as xsd:token and xsd:string. The idea is to make our language useable (specifically enumerations and key/keyref) even without a separate datatyping vocabulary. The only thing lacking relative to XML 1.0 apart from legacy cruft (eg ENTITY) would be the ability to restrict a string to name characters.
James
