RE: <restriction> - Bad Kleene Notation

Hi Tim,

     You wrote (in part)

                 However, it does appear that the use of the Kleene
Operator in the XML Schema document is incorrect.  I don't
believe it is valid to define "conditions" that change the
definition of the "*" operator.  Since the "*" operator
is _defined_ to mean "zero or more occurrences of", it
is erroneous to then (verbally) qualify the operator in other
sections of the document.  This is similar to changing
the meaning of the mathematical "+" operator.  I believe the
production rule for the definition of the <restriction>
statement is invalid and the editors should correct it.
(Reference: XML Schema Part 1: Structures - 3.4.2 XML
Representation of Complex Type Definitions
http://www.w3.org/TR/2001/REC-xmlschema-1-20010502/#element-simpleContent::restriction)


     Although you are correct, I don't think it's not very practical to 
express this any other way.

     For example, consider a regular set of strings consisting of any 
number of the letters a and b, but at most one instance of each of the 
letters c and d.  One way of specifying this is as follows:

  (a|b)*((c?(a|b)*d?)|(d?(a|b)*c?))(a|b)*

It's much easier to express by saying it has the following form:

  (a|b|c|d)*

with the added restriction that that there is at most one instance of each 
of c and d.

     We (meaning humans) abuse notation to simplify things in many other 
situations.  For instance, the syntax of a programming language is often 
described using a context-free grammar, but few programming languages are 
actually context-free languages.  For instance, there are often 
requirements that the use of a name be preceded by a declaration of the 
name (as a variable, say).  That can't be expressed using a context-free 
grammar, but we do it anyway, with a prose or semantic description of the 
added restriction.

Thanks,

Henry
------------------------------------------------------------------------
Henry Zongaro      XML Parsers development
IBM SWS Toronto Lab   Tie Line 969-6044;  Phone (905) 413-6044
mailto:zongaro@ca.ibm.com