This is an archived snapshot of W3C's public bugzilla bug tracker, decommissioned in April 2019. Please see the home page for more details.
The specification for "redefines" in 4.2.2 states: "In order to provide some support for evolution and versioning, it is possible to incorporate components corresponding to a schema document with modifications. The modifications have a pervasive impact, that is, only the redefined components are used, even when referenced from other incorporated components, whether redefined themselves or not." There are many difficulties in understanding this text, centering around the word "pervasive". The use of "even when" suggests an example rather than a definitive statement of what "pervasive" means. However, the clear implication is that any schema component A that refers to a schema component B that has been redefined as B' is taken as referring to B'. However, this intepretation doesn't work for attribute groups and model groups, because no schema component ever contains a reference to an attribute group definition or a model group definition. Perhaps this is why we read in the note "The pervasive impact of redefinition reinforces the need for implementations to adopt some form of lazy or 'just-in-time' approach to component construction". The implication of this note is presumably that conversion of the information in a schema document to a schema component is done as late as possible. Presumably once this conversion has been done, subsequent redefinition of components will have no effect (at any rate in the case of model groups and attribute groups, but perhaps in other cases too). In any case, the phrase "referenced from other incorporated components" seems wrong. It is references from schema documents that are redirected, not references from schema components. Once the schema components have been built, there is no reference to any attribute group or model group definition, so no redirection is possible.
On its telcon today, the Working Group discussed this and other issues related to schema composition and concluded (not without some pangs of regret) that for scheduling reasons it is not feasible for us to resolve them before we go to Last Call. The WG continues to believe that this issue is important and should be resolved; our decision to close the issue for now reflects the fact that we have not been able to reach consensus on how to resolve the problems related to composition, coupled with the observation that for the most part, users are able to avoid the problems by avoiding problematic constructs in their schema documents. The cost of delaying XSDL 1.1 for these issues is thus high, while the practical benefit for users is not as high as it might be. The Working Group continues to hope that this issue can be resolved at some point in the future. Accordingly I am closing this issue with a disposition of LATER.