Philippe,

how about this:
  - The grammar based languages provide a regular grammar
  - Schematron provides boolean assertions that can express
  constraints between distant parts of the document tree
  - CLiX (http://www.systemwire.com/xlinkit/clix.html) provides first
  order assertions that can express constraints between distant parts
  of the document tree, or between multiple documents our xlinkit
  implementation
  - Programming languages provide turing completeness on top of that.

There is a hierarchy of expressiveness here: schemas -> schematron ->
CLiX -> Programing Language.

At the last step, you drop off the declarative level and start
hacking.

Relationships between multiple data sources are orthogonal to that:
you can still classify approaches by expressiveness between multiple
data sources.

Christian

> hi,

> An XML Schema is the expression of some assertions expected on an XML
> document class. Assertions on XML documents ensure that applications
> will process them without causing faults. Expressing assertions with
> schemas ensure that applications developpers will spend most of their
> time in designing data process and few of their time in controlling them.

> Any schema technology is designed to cover numbers of assertions 
> expressed. However, if think that the existing schema technologies (DTD,
> W3C XML Schema, Relax NG, Schematron) can't express some constraint types.

> I have some ideas of what these constraint types might be, but I'd like
> to know, helped by your own experiences, which kind of useful assertions
> can't be expressed with these schema technologies (so frustrating).

> For example :
> -i'd like that the number of occurrences of an element is equal to a
> value of a given attribute (Schematron could do that)
> -i'd like that the value of a given attribute exists in my RDBMS
> -etc...

> Thanks in advance