XML News from Monday, July 7, 2008

The W3C XML Schema Working Group has posted new last call working drafta of XML Schema 1.1 Part 1: Structures and XML Schema Definition Language (XSD) 1.1 Part 2: Datatypes According to the structures draft,

XSD 1.1 retains all the essential features of XSD 1.0, but adds several new features to support functionality requested by users, fixes many errors in XSD 1.0, and clarifies wording.

This draft was published on 20 June 2008. The previous working draft of 30 August 2007 was a Last-Call Working Draft which elicited numerous comments and suggestions for improvements. All substantive issues have now been resolved, although some editorial issues remain open. The major revisions since the previous draft include the following:

• The minimal subset of XPath which processors were required to support for assertions has been eliminated; processors must support all of XPath.
• A new wildcard keyword ##definedSibling has been added to allow a wildcard to match any element except one mentioned explicitly elsewhere in the current content model.
• The definitions of must and ·error· have been revised to require that processors detect and report errors (although the quality and level of detail of the error messages are not constrained).
• An <override> element has been defined to allow the declarations or definitions of specified components in other schema documents to be overridden.
• The <redefine> element has been ·deprecated·.
• XML Representation Constraints no longer refer to the component level; they can now be checked for schema documents in isolation.
• Numerous editorial changes and clarifications have been made and numerous small errors corrected.

In the datatypes spec,

The previous working draft of 17 February 2006 was a Last-Call Working Draft which elicited numerous comments and suggestions for improvements. All substantive issues have now been resolved, although some editorial issues remain open. Changes since the previous public Working Draft include the following

• Support has been added for assertions on simple type definitions, analogous to those allowed by [XSD 1.1 Part 1: Structures] for complex type definitions.
• The requirements of conformance have been clarified in various ways.
  A distinction is now made between ·implementation-defined· and ·implementation-dependent· features, and a list of such features is provided in Implementation-defined and implementation-dependent features (normative) (§H).
  Requirements imposed on host languages which use or incorporate the datatypes defined by this specification are defined.
  The definitions of must, must not, and ·error· have been changed to specify that processors must detect and report errors in schemas and schema documents (although the quality and level of detail in the error report is not constrained).
• Conforming implementations may now support ·primitive· datatypes and facets in addition to those defined here.
• A number of syntactic and semantic errors in some of the regular expressions given to describe the lexical spaces of the ·primitive· datatypes have been corrected.
  The character sequence '+INF' has been added to the lexical spaces of float and double.
• Parts of the grammar of regular expressions given in Regular Expressions (§G) has been revised for clarity. The set of legal regular expressions has not been changed since the previous Working Draft.
• The lexical mapping of the QName datatype, in particular its dependence on the namespace bindings in scope at the place where the ·literal· appears, has been clarified.
• The characterization of ·lexical mappings· has been revised to say more clearly when they are functions and when they are not, and when (in the ·special· datatypes) there are values in the ·value space· not mapped to by any members of the ·lexical space·.
• The nature of equality and identity of lists has been clarified.
• Enumerations, identity constraints, and value constraints now use equality-based comparisons, not identity-based comparisons, in cases where there is a difference between identity and equality.
• The mutual relations of lists and unions have been clarified, in particular the restrictions on what kinds of datatypes may appear as the ·item type· of a list or among the ·member types· of a union.
• Unions with no member types (and thus with empty ·value space· and ·lexical space·) are now explicitly allowed.
• Cycles in the definitions of ·unions· and in the derivation of simple types are now explicitly forbidden.
• Errors have been corrected in the description of the canonical mapping of decimal.
• A variety of clarifications have been introduced connected with the terms "finite-length", "actual value", "type", "datatype", "·special value·", and others.
• A number of minor errors and obscurities have been fixed.

Comments are due by September 12.