Re: Closure for XML Schemas

Murali Mani <mani@CS.UCLA.EDU> writes:

> My previous question to the schema WG was about CFGs appearing in the
> content model. You had replied that it was handled in a way similar to
> what RELAX does -- by not allowing recursion in model groups.
> 
> I had another question now -- will there be a definition of closure
> properties under boolean set operations, such as union, intersection and
> difference for XML Schema? I believe, XML Schema will not be closed under
> union and difference. It should not be surprising that most query
> operations also will not be closed under XML Schema. I believe one of the
> usage scenarios of XML Schema is for query?

It's clear that this isn't possible without a major change to the XML
Schema design, which depends on the Unique Attribution constraint in a
number of ways.  This means that the union of two types may be
undefined, e.g.

(a?,b*) U (b*,c)

> And databases do not work without closure. Also, I do not think we can
> obtain closure by restrictions on XML Schema. XML Schema is not closed
> because it is much restricted and less expressive than, say RELAX.

That restriction buys both implementation simplicity and expressive
power in other areas (annotation, keys, etc.) which were judged very
important.

> Is this something which XML Schema will consider?, in such a case, I find
> it better that for some purposes, we use a closed schema language. I
> believe RELAX will be closed under all boolean set operations as well as
> query operations (from the properties of regular tree languages). 
> Therefore will it be possible for the Schema WG to combine forces with
> RELAX, and define something when we require such closure properties. From
> where I stand, it appears that RELAX is not being given any consideration,
> which I believe it deserves.

We considered this issue long and hard, but came to a different
conclusion than you evidently have.  

> I would appreciate the views of the Schema WG on this.

What you've got is my personal views:  I don't speak for the WG except 
when empowered by explicit WG decisions.

ht
-- 
  Henry S. Thompson, HCRC Language Technology Group, University of Edinburgh
          W3C Fellow 1999--2001, part-time member of W3C Team
     2 Buccleuch Place, Edinburgh EH8 9LW, SCOTLAND -- (44) 131 650-4440
	    Fax: (44) 131 650-4587, e-mail: ht@cogsci.ed.ac.uk
		     URL: http://www.ltg.ed.ac.uk/~ht/

Received on Friday, 8 December 2000 06:29:40 UTC