XML Schema 2nd Edition

Related link: http://lists.xml.org/archives/xml-dev/200403/msg00372.html

The W3C XML Schema Working Group has published the
second edition of the three parts to XML Schema.
This is a housekeeping edition.

The changes of note in the XML Schemas Second Edition seem to be:

• Some examples in the Primer are simplified,
• The Schema for Schemas is more open now, allowing attributes from other namespaces everywhere,
• The Structures spec tidies up some loose ends concerning substitution groups,
• The Datatypes spec tidies up a bug with dates (potential for incompatibility could be a gotcha for some existing users) and allows the pattern facet in more places, and
• The definitions, status and allowed uses of the ur-types anyType and anySimpleType have been rejigged: they are considered real types now, not just theoretical apparatus; apparantly there were some edge cases for lists or unions that made this necessary. Maybe it will help some future transition to XQuery's type hierarchy too?

Should you go rushing out to read it all? Well, if you enjoyed reading it the first time around, all your favourite passages will still be there! I have heard that the film rights have being negotiated for an epic trilogy, but the producers are not sure whether to make it in the style of the Matrix (programmer wakes up in a nightmare world, utterly impenetrable, sponsored by Oracle), LOTR (enormous scale, but only the innocent can handle something so powerful without being corrupted) or Harry Potter (problems solving with a magic wand, but full of goblins).

The Working Group also has a call out for requests for an XML Schemas 1.1; my prediction for XML Schemas 1.1 will be XML Schemas 1.0 harmonized with the XQuery type hierarchy, and with document-oriented SGML-isms like NOTATION, ENTITY* removed: slightly more DBMS-oriented and even less publishing-oriented. I am not holding my breath that the big boys will be keen on anything more disruptive than that, but in the other hand as users get enough experience to sort out the wheat from the chaff we might get some widespread support for SGML-to-XML-style refactoring eventually.

Anyway, congratulations to Henry, Michael, Dave and the Working Group.