XDI metaschema and Roger Costello's "Extreme Extensibility"

["Drummond Reed" <drummond.reed@cordance.net>]

XDI TC Members and Observers:

I had the action item last week to create an XML example of how the proposed
XDI metaschema would be used in order to address several issues, including
resource typing, addressability, link structure, etc.

My workload of this last week has prevented me from creating the type of
example that I would like. However, in thinking through the points that this
example XML document would illustrate, and thinking about the recent posts
to the link regarding these issues, it occurred to me that there are even
higher-level design issues involved here.

I believe these issues are fundamentally those involved with designing an
XML metaschema vs. an XML schema. A XML metaschema is not designed to
replace existing XML schemas. It's designed to provide a new way to describe
the structure of XML data (a schema). The W3C XSD spec is a metaschema. The
RDF spec is a metaschema. I believe what we're designing a different type of
metaschema that is optimized for cross-domain data sharing, linking, and
synchronization. As such, there are certain design principles for such a
metaschema that are different than those for ordinary XML schemas.

This led me back to an article that provided the inspiration for the
development of the initial proposed XDI metaschema. Called "Extreme
Extensibility" by Roger Costello, it's posted on the Xfront site at:

	http://www.xfront.com/eXtreme-eXtensibility.html

Roger's article has several very clear XML examples of the XML design
pattern he recommends for extreme extensibility. If possible, please review
this article prior to today's XDI TC call, as the examples there are a good
stand-in for several of the points I was planning to illustrate with my
example. (Which I am still working on, but likely will not have ready by
today's call.)

=Drummond