Re: drums2?

> You need therefore some foolproof way to distinguish between an encoded
> local-part and an unencoded one. This means placing some restriction on
> local-parts as presently defined.
you do the same thing that IDN did:

- you pick an encoding that is extremely unlikely to have been used in 
  a real local-part.

- you *always* use the encoded form on the wire.

it doesn't require a change to the message format.

Keith