> As an SGML publsihing guy I found this to be a horrifying situation but 
> over the years working with the Insurance industry, I've come to accept 
> the need for this. Now the question is how do you support it and with 
> which features? The one group that uses xsd:any is not interested in 
> validatiing the extensions, the other group is. We are not happy that 
> redefine modifies the primary namespace, but we have to live with that 
> outcome to get the functionality we are looking for "out of te box" with 
> standard XML schema tools.

This is exactly the thinking in the MISMO group. Alongside our allowance 
of <redefine> is a guideline on how extensions to the core schemas 
should be done. Attributes and elements from other namespaces may be 
added to the content models of existing elements, but nothing can be 
taken away. Because the extensions must be namespaced processing the 
core elements and attributes is rigid and tools that were designed to 
operate on the core schemas can generally be re-used in the extended 
versions.

Cheers,
Jeff Rafter

From dvint@d... Sun Jul 10 21:04:06 2005
Received