Re: [Uta] REQUIRETLS: another SMTP TLS mechanism

[Jim Fenton <[e-mail]>]

On 03/25/2016 07:24 AM, Jeremy Harris wrote:
> On 25/03/16 02:12, Jim Fenton wrote:
>>  draft-fenton-smtp-require-tls-01
>> The idea here is that REQUIRETLS allows the SMTP client to override the
>> default "deliver even if you can't do it securely" behavior of SMTP. The
>> philosophy is that the sender of the message (SMTP client) is in the
>> best position to know if a given message should only be sent via TLS,
>> either based on some information it has about the sensitivity of the
>> message or based on the client's local policy.
> - The draft does not mention alias-style forwarding done by an MTA;
>   perhaps it could?  A 1-1 alias would seems to be easily covered,
>   but 1-to-many (mail-exploder) aliases may need more thought.

REQUIRETLS is about satisfying the sender's expectations about how
onward transmission of the message takes place. As a sender, my
expectations wouldn't change if an MTA forwarded the message to more
than one recipient, so I would want REQUIRETLS to be applied to all of
the onward forwards. I agree, this could be more explicit.
>
> - I assume that mailing-list forwarding is not intended to be covered.
>   Perhaps that should be explicit.

We had a little discussion about mailing lists and REQUIRETLS on the
ietf-smtp list. Mailing lists (as distinct from mail
forwarders/exploders, above) originate a new message from the mailing
list. So I would expect REQUIRETLS from a mailing list to be at the
option of the list operator, although it would be a good practice in
many cases to follow the REQUIRETLS practices of the incoming message. 
Agree on making this more explicit, too.

-Jim