xml-dev - RE: XML Schemas: Best Practices

["Bullard, Claude L (Len)" <clbullar@ingr.com>]

Staring down the slippery slope....

From: Roger L. Costello [mailto:costello@mitre.org]

>[1] Do you agree that, by themselves, schemas define just syntax and not
>semantics?

>>Yes.  OTOH, an abstract node can't be instanced.  This communicates a 
semantic.  There is the semantic of THE schema language and the semantic 
to be sustained by use of A schema language.  But we do not know the 
meaning of the abstract node semantic unless we:

1.  Read the record of authority.
2.  Observe the behavior of an application claiming to sustain  
    the semantic described in the record of authority and attest 
    to the correctness of that behavior.

>[2] A big question ... what is "semantics"?  

It is a point of view.  It is organized by contract or 
by behavior.   Something means what you say it means.  
Universality of meaning is an illusion sustained behaviorally and 
communicated by context organized by contracted representation.  It 
is more precise to refer to a schema as an application vocabulary 
over simply a vocabulary.  To sustain a semantic is to ensure the 
predictability of a behavior.

>>[4] How do we "add semantics" to a schema document?

By contract (document, comment, etc) or by reference.  
The means by which the semantic is defined is outside the scope of 
the XML Schema language and should remain so.  The point 
at which one begins to add the semantic definition (say, 
executable instructions) to the schema language, is the 
time to abandon XML schemas and use a well-practiced programming 
language.