[mail-vet-discuss] SHOULD the header be signed?

[Murray S. Kucherawy]

Charles Lindsey wrote:
> I might not go as SHOULD, but certainly the practice should be 
> encouraged in suitable cases. These include
>
> 1. where the mail is to be sent further using SMTP (whether within the 
> final delivery boundary or not - note that such boundaries are not 
> always clearly recignised, even within their supposed borders).
Does this draft need to say "to protect this header, you should sign 
it"?  That seems to be something generally true and not specific to this 
proposal.
> But, as a corollarly, it should be stated that these headers SHOULD 
> NOT be removed at boundaries in cases where they are covered by such a 
> signature (I probably mean a signature that verifies correctly).
The final paragraph of "Removing The Header Field" discusses that issue 
already, though not in a normative sense.  Is that really necessary?