Re: Counter-proposal for XML Schema: XML Notation Schemas

In his note entitled "XML Notation Schemas", Rick pointed out 
"two severely broken non-conformances", namely:

- supporting non-well-formed XML 1.0; 
- treating namespace URIs as schema names. 

I believe that Rick is right.

One of the ideas he suggested is the use of XSL patterns for describing schemata.

>Using XSL as a Structure Validation Language, which characterizes document 
>validation as as stylesheet issue and points out that XSL's pattern language 
>allows validation on very different criteria to SGML's content model approach; 

If we allow non-terminals in the schema language, we can construct a schema 
from an XSL pattern (probably, with some restriction).  A document is valid 
against this constructed schema if and only if some element in this document 
matches the given pattern.

Formally speaking, a pattern can be captured by a pair of tree regular 
expression (conditions on subordinates) and pointed tree regular expression 
(conditions on non-subordinates such as superiors, siblings, etc).  Kleene's 
theorem ensures the construction of a tree regular grammar from a tree regular 
expression.  My work (PODP'96) allows for the construction of unambiguous 
non-determinstic tree automaton from a pointed tree regular expression.  
We then remove matching states from this unambiguous automaton.  By applying,
subset construction and negation, we can construct a tree regular grammar 
that mimicks the pointed tree regular expression.

Makoto
 
Fuji Xerox Information Systems
 
Tel: +81-44-812-7230   Fax: +81-44-812-7231
E-mail: murata@apsdc.ksp.fujixerox.co.jp