Re: [relax-ng] close schema language

[MURATA Makoto <EB2M-MRT@asahi-net.or.jp>]

Michael Fitzgerald wrote:

> I was wondering what people think of these comments on closed schema
> languages on XML-DEV.

I see some advantages in assertion-based schema languages such as Schematron.  
But I have always wondered why Rick Jelliffe opposes to other languages.  
What he calls "closed" is not a hangover from DBMS.  Does he really think 
DocBook and TEI are such hangovers?  "Closed" content models are sometimes 
(or often) required.

It is true that handling of foreign namespaces is still a challenging issue.  
RELAX NG provides name classes.  We have to examine if they are powerful enough 
for capturing foreign namespaces and, if necessary, persue other approaches 
(e.g., XSLT transformation followed by validation).

BTW, appendix A of the compatibility spec provides an interesting schema.  It is 
completely different from the schema in the RELAX NG specfication, but is intended 
to be used in conjunction with it.  Both schemas heavily use name classes.

Cheers,

Makoto