This is an archived snapshot of W3C's public bugzilla bug tracker, decommissioned in April 2019. Please see the home page for more details.
QT approved comment: In 2.6.1.3, the definitions of "transitive membership" et seq make the unstated assumption that the relation between a union type and its member types is acyclic.
Thank you; yes, the discussion does assume that the relation between a union and its member types is acyclic (or, equivalently, that no union is a member of its own transitive membership). This is explicitly required by the Schema Representation Constraint: Simple Type Definition Representation OK in section 3.16.3 of part 1 (Structures) of XSDL 1.1. It's not clear to me how Structures section 3.16 and Datatypes section 4.1 have gotten out of synch again, after a long and trying effort to reconcile them. But it seems clear that they need to be reconciled once more.
The WG discussed this issue (and in particular the alignment of Structures and Datatypes) at its call of 26 October 2007. We concluded that the goal of the earlier alignment work was not to make the relevant sections of Structures and Datatypes effectively identical, but only to ensure that material appearing in both specs should be the same in both. This includes the tableaux and some textual descriptions. The goal had not been (we believe) to require that every constraint in these sections of Structures also appear in Datatypes, and vice versa -- only the ones that 'belong' in the particular spec. (The notion of belonging is of course slightly fuzzy.) The WG instructed the editors to compare the relevant parts of the two specs and make recommendations for duplicating or moving text as appropriate. With regard to the particular issue raised here, the WG instructed the editors to prepare a wording proposal to capture, in Datatypes, the constraint that no union be a member of its own transitive membership. So the status of this bug is changed from needsAgreement to needsDrafting.
A wording proposal intended to resolve this issue is at http://www.w3.org/XML/Group/2004/06/xmlschema-2/datatypes.b3235.html (member-only link).
The wording proposal mentioned in comment #3 was adopted by the WG at today's call; with that, I am marking the issue resolved. Michael, as the originator of the issue and as our contact wtih the QT working groups, would you please report on the resolution to QT and indicate by closing the issue that they are happy with the resolution (well, satisfied if not happy), or by reopening it that they are not satisfied? If we don't hear from you within the next two weeks, we expect to assume that silence implies consent.