xml-dev - Re: Schema concepts (was Re: W3C public lists)

[Stefan Haustein <stefan.haustein@trantor.de>]

Michael Rys wrote:
> 
> In addition, I would like to point out that in Object Data Models (including
> the later releases of ODMG), there is a difference between structural type
> lattices (classes in C++, aka type inheritance) and semantical collection
> hierarchies (collection of objects that satisfy certain conditions that are
> either necessary and sufficient or necessary, aka as class subsumption). A
> common condition on such a semantical collection is, that all members are of
> the same base type (the so called member type). This is exactly how you can
> interpret the distinction between type (structural description) and element
> (semantical role).

The point in OOP is class subsumption, not structural 
derivation. You cannot reuse the structure without class 
subsumption. Why is that different in XML schema? I do
not see the reason. If you are used to OOP (many people are) 
the XML schema specs are confusing. It is possible to add
everything you want without violationg basic OOP principles, 
and without forcing people to learn yet another 
paradigm. For example, in JAVA, you have interfaces for 
the semantical roles without structure, but you do not need 
to use interfaces in your own class hierarchies if you think 
you do not need them. 

Best regards

Stefan Haustein

-- 
SAX-based access to WBXML and WML: http://www.trantor.de/wbxml
XML pull parser: http://www.trantor.de/xml