This is an archived snapshot of W3C's public bugzilla bug tracker, decommissioned in April 2019. Please see the home page for more details.
At the end of section 5.3: "In the case of element information items, processors MAY choose to continue ·assessment·: see ·lax assessment·." But in the definition of "lax assessment" in 3.4.4: "If the item cannot be ·strictly assessed·, ...the element information item's schema validity MUST be laxly assessed ..." Note the difference between MAY and MUST. A couple of issues: - Since the intention is to require "lax assessment", seems 5.3 needs to be updated to also use "MUST". - The current definition of "lax assessment" doesn't seem to cover the case mentioned in 5.3.
Concerning Validation Rule: Schema-Validity Assessment (Element) From the minutes... RESOLUTION: we have phase one agreement to 1) attempt to reword to say "An element information item's schema-validity is strictly assessed if and only of all of the following are true": and 2) look at all instances in the spec of the words "strictly assessed" and dependent terms to make sure we are not causing trouble. MSM: We should distinguish two things. 1) what does the phrase "lax assessment mean" - it means use element locally valid type with xs:anytype, assessing children recursively lax, with certain PSVI consequences 2) separate from the definition of lax assessment, there should be statements that in some conditions where strict can't be done, lax can be. MSM: Part of the problem is making the conditions under which to do it part of the definition. proposal: rework definition and the "if X then do lax assessment" rules to make the separation. RESOLUTION: For issue 5412, we will rework the definition of lax assessment to make the separation between definition of the term vs. when lax should be attempted. Bug will get keyword "needsdrafting"
The initial point (discrepancy between MAY in 5.3 and MUST in 3.4.4) has been resolved: 5.3 now uses MUST, not MAY. (The change was part of the wording proposal for bug 3251.) The remaining points appear to be clarifications rather than substantive changes, so I am marking this bug report editorial. (This means it may be addressed after, rather than before, the spec goes to Last Call.)
During its 2009-03-20 telecon, the schema WG adopted a proposal to address this issue. The changes is to update the "lax assessment" definition to also cover the "missing component" case. With this change (along with comment #2), the WG believes that the issue raised in this bug report is fully addressed. I'm marking this RESOLVED accordingly. As the reporter of this bug, I will also close it shortly.