Re: Order indicators in the .xsd file

On Oct 17, 2012, at 8:39 AM, George Cristian Bina wrote:

> Expression every possible permutation does not generally work as you will get a lot of errors from the unique particle attribution rule that will not allow you to say for example
> 	(a, b, x) | (a, b, y)
> forcing you to write
> 	(a, b, (x | y))
> instead.
> 
> However, for a complex model getting the right expression is not easy and sometimes it may be impossible.

I agree that it is often difficult, and the result is almost never
easy to read and understand.  But I do not believe it is ever 
impossible.  

There is a definable class of regular languages for 
which deterministic content models are impossible, and the
problem of interleaving a number of elements with different
occurrence constraints does not look at first glance as if it had
the required properties.  (But I have not attempted a formal
proof of this proposition and I may be wrong.)

-- 
****************************************************************
* C. M. Sperberg-McQueen, Black Mesa Technologies LLC
* http://www.blackmesatech.com 
* http://cmsmcq.com/mib                 
* http://balisage.net
****************************************************************