This is an archived snapshot of W3C's public bugzilla bug tracker, decommissioned in April 2019. Please see the home page for more details.
When assertions are evaluated, a data model instance must be constructed for the evaluation. The Structures spec needs to explain how that data model instance is created and in particular make clear what type labels it has. The WG has discussed this at length in the context of a more general proposal for 'tree trimming', and agreed at the meeting of Friday 23 March to adopt the 'Moses' proposal, which involves validating as much of the element in question as is feasible without checking its assertions, building a PSVI to reflect that partial validation, and using the resulting data model instance. There is one point of concern about this procedure: when the element in question has a complex type with simple content, it seems unfortunate that the value of the element itself is not typed. So the editors were asked to see if a simple way could be found to type the value. (This addition to the Moses proposal has acquired the soubriquet "Moses supposes".) The data model construction rules should also be applicable to XPath evaluation in identity constraints and in conditional type assignment, should the WG adopt a conditional-type proposal.
I'm changing the keyword to 'needsDrafting' to reflect the current state of this topic more accurately.
Updates from 207-05-25 telecon: - Don't separate out XDM construction rules from Assertion now. May or may not do it under CTA. - Try to handle precisionDecimal in a way consistent to QT's suggestions.
The WG reached consensus, in a series of calls in May and June 2007, on rules for evaluating XPath expressions in the context of schema-validity assessment, including both specification of how to type the XDM instance or leave it untyped (bug 4416) and specification of the static and dynamic contexts (bug 4419). A wording proposal reflecting that consensus is at http://www.w3.org/XML/Group/2004/06/xmlschema-1/structures.b4419.200707.html (member-only link). Accordingly, I'm changing the status of this issue to needsReview.
The wording proposal mentioned in comment #3 was discussed and adopted by the WG during our call of 3 August 2007. With that decision, this issue has been successfully resolved, and I am setting its status accordingly.