I think that Dan has a point that if a reliable alternative mechanism is
available for tracing message origin, then it shouldn't be necessary to
watermark the content. But I do wonder if the available IPP mechanisms
would be sufficient in all cases (e.g. GSTN/G3fax to IPP/FAX onramp).
On the subject of timestamp, I think the only one we have to be concerned
with in a wire protocol is the date/time of sending. All other timestamps
are provided by the receiver and hence are local implementation issues.