Yes, I am not surprised. Chris makes this assertion often and I have made him a promise that I would dispute it every time he makes it ;-). We have agreed to make the spec work with or without a CPA. This presents the potential conflict of having MessageHeader element conflict with CPA values. To resolve this conflict, we agreed not to have header values override CPA values (if one exists). This does not indicate a need to remove all these potentially conflicting values from the headers. Regards, David.