Re: [xml-dev] Some random noise on rational type systems for XML

Rick Jelliffe wrote:
> 
> ...
> 
> I guess among the possibilities are that value spaces and lexical spaces
> form distinct hierarchies, so a data type should be a mapping from one or more
> lexical types to one or more value types.
> 

i suggest that you can phrase that somewhat stronger, that value spaces and
lexical spaces which form distinct hierarchies are the _only_ general form.
other forms are special cases only. and the schema "data types" are really
function types.

yes. finally.

i haven't read other people's code in his regard, but were i to, i suspect i
would find that i am not the only one who reifies schema types as classes
(which implies specialized methods) and/or globally named functions which
translate between appropriately constrained lexical and internal data domains.

now if w3c specs whould only use terminology which made it possible to express
this clearly.

...