The refinement consisting in allowing conforming consumers to stay silent (at their will) in front of documents produced by non-conforming producers (that would violate the rule that no element should be identified by a ProcessContent attribute and have an xml:lang or xml:space attribute) only complexifies the specification, for dubious (and undocumented) benefits. It would be much clearer and simpler to state that conformant documents should not include elements that are identified by a ProcessContent attribute and have an xml:lang or xml:space attribute.

Replace by a clear statement about conformant documents, or else provide a rationale for the unneeded sophistication introduced here.

page 15, lines 27-31 Part 5, Section 9.1.2 ProcessContent Attribute

te

Proposed Disposition of DIS 29500 Comment FR-0572 (Modified: 2008-01-03) Agreed, the text should not place limitations on non-conformant input; the following change will be made to Part 5, §9.1.2, page 15, lines 27­31: Markup producers shall not generate an element that has an xml:lang or xml:space attribute if that element is identified by a ProcessContent attribute value. Markup consumers that ignore an element that has an xml:lang or xml:space attribute and is also identified by a ProcessContent attribute value shall generate an error. Markup consumers that encounter a non-ignored element that has an xml:lang or xml:space attribute and is also identified by a ProcessContent attribute value might generate an error.