Proposed fix for DR 10-0016 ("MCE: Core Concepts")
alexb at griffinbrown.co.uk
Tue Jul 6 22:55:48 CEST 2010
> But isn't this the essence of Alex's complaint - that Microsoft Office
> 2010 adds new attributes from foreign namespaces to known elements? (a
> good example of this is the new attribute on the C-element in
> SpreadsheetML (I have forgot the name of it - is it dXy or something?) -
> thereby invalidating the original element content wrt the original schema
Yes - but I'm prepared to be overruled on this.
My original thought was that the "original" OOXML Namespaces, as governed by schemas, were guaranteeing a certain form of markup and that the subsumption mechanism of MCE was intended to get implementers to preserve the sanctity of that markup, by forcing them to innovate only in all-new Namespaces.
The interactions between the conformance statements of Parts 1 & 4, the mechanisms of MCE, and the unstated requirements of MCE mean that currently we have an urgent problem in that what is described is not implementable in any testable conformant way, and even the bits that are described have defeated any attempt (that I have seen) to understand them. One way or another, we need a solution and my proposed text is certainly one way to do this ...
This email has been scanned by the MessageLabs Email Security System.
For more information please visit http://www.messagelabs.com/email
More information about the sc34wg4