This is an archived snapshot of W3C's public bugzilla bug tracker, decommissioned in April 2019. Please see the home page for more details.
Assume that in an instance document, element "e" matches a lax wildcard and there is no global element declaration for "e" and there is an xsi:type specified on it. How *valid* does the xsi:type need to be? Not a qname: <e xsi:type="a b c d"/> Not resolve to a type: <e xsi:type="randomNN:randomName"/> This doesn't seem to be covered by Schema-Validity Assessment (Element) in 3.3.4. Does this mean it's valid (and should be treated as if xsi:type is not there)? Is this the intention? See this exchange (member-only) http://lists.w3.org/Archives/Member/w3c-xml-schema-ig/2005Mar/0098.html
See also bug 2225. If 2225 is resolved by making it clear that xsi attributes must be validated wherever they occur on elements that aren't being skipped, then at least part of this issue will also be resolved. Bug 2225 does not resolve (as far as I can tell) the question "what happens if there is no element declaration, and the QName value of xsi:type does not resolve to a type?"
The Working Group adopted a wording proposal to resolve this issue, during our meeting of 23 March. The upshot is that a new rule is added to Validation Rule: Attribute Locally Valid, reading: 5. If the declaration is for xsi:type (Attribute Declaration for the 'type' attribute (§3.2.7)), then the item's ·actual value· ·resolves· to a type definition. Accordingly, I'm marking this 'closed'. Sandy, I believe that you originated this issue; if so, please shift it from RESOLVED to CLOSED to indicate your acceptance of the proposal you drafted. If you were channeling someone else, please let them know. Thanks.