equivClass: common ancestor type (revisited)

In your slides (http://www.simonstl.com/articles/schemas/schema13.htm), you
said:

Equivalence    XML Schemas lets you declare different types equivalent with
the equivClass attribute. This makes them more interchangeable, somewhat
like objects of a common parent type, though no common parent is required.

I think that would be most peoples expectation, but that seems to disagree
with section 2.2.2.2.  I discussed this in
http://lists.w3.org/Archives/Public/www-xml-schema-comments/2000AprJun/0140.
html arguing that the common ancestor type requirements seems to be an
unnecessary and unbeneficial restriction.

Maybe you know something about the current state that I don't, but this
seems to be one issue that I raised that didn't get a last call issue number
that probably should have unless it is buried in some other issue.

---------
A six-month (or one year) timeframe?    Software developers may face a
heavier learning curve - they get to implement all of this and make sure
it's correct before shipping it. So far, there hasn't been any work on a
conformance suite, and most schema implementations are at least one draft
behind.

Henry Thompson, Mary Brady of NIST, David Brownell, the Xerces-J mailing
list had some threads on this issue in the last few weeks.  The testcase.dtd
and David Brownell's XML conformance harness should only need minor
enhancement to support schema conformance testing.  Henry was supposed to
post a message on xerces-j after he had thought things through.  Probably
the most expedient would be to set up a SourceForge project, though Apache,
W3C or OASIS might be the better official home.

See http://xml-archive.webweaving.org/xml-archive-xerces-j-dev/1004.html

Received on Monday, 3 July 2000 12:57:41 UTC