schemas and transforms, was: Path 1.5? (was typing and markup)

David,

> 
> The claim here was that one could take a transform (in Xquery or perhaps
> XSLT) and typecheck that it forms a function from the set of documents
> of type specified in schema-1 (say docbook) to the set of documents of
> type specified in schema-2 (say XHTML).
> 

Yes I think that is correct. With a correctly constructed
formal semantics (i.e. assuming XQuery FS works :-), 
that is indeed what I expect would be a benefit, perhaps
the major benefit.

I would go as far as to make that a test case or if I
were on the WG, to make that a requirement, 
perhaps it is.

One of the reasons that I have become interested in such 
formalisms, is that given properly written schemas or more
generally ontologies, I would like _software_ to write me
these transforms -- this is almost exactly what the 
"Semantic Web" means to me. I think its possible.

Jonathan