+1
Doug Bunting wrote:
> David,
>
> This doesn't have the appearance of a non-normative note. And, as Marty has
> mentioned, it may be worthwhile limiting this to a disclaimer:
> "persistDuration doesn't have anything to do with message order
> implementation." Some explanation might be sensible.
>
> This goes the other way, recommending a specific implementation.
>
> I recommended, in my comments on the schema and the second half, removal of
> the messageOrderSemantics attribute and requiring SequenceNumber in
> MessageOrder. The current module is more complicated than necessary,
> seemingly allowing a sender to provide information the receiver can randomly
> choose to use. It raises a red flag to see text like "X element MUST appear
> when Y attribute has the value Z".
>
> thanx,
> doug
>
>