Re: draft-kucherawy-greylisting-bcp

[Steve Atkins <[e-mail]>]

On Oct 26, 2011, at 8:33 AM, Pete Resnick wrote:

> 
> On 10/26/11 10:18 AM, Dave CROCKER wrote:
> 
>> On 10/26/2011 5:08 PM, Pete Resnick wrote:
>> 
>>> If it can be deployed and you
>>> get get incremental implementation experience, it should be on the Standards
>>> Track; it shouldn't be a second class citizen of being a BCP.
>> 
>> Greylisting is not a protocol, according to any definition I know.
>> 
>> There is no way to test "interoperability".
> 
> Rubbish. I can assure you that if someone implemented greylisting in such a way that "legitimate" mail stopped being delivered, because they set the greylist timeout too long or because they used a status code that everyone choked on or some other such thing that was underspecified in the RFC, people would say that the implemenation was not interoperating in the mailing transport/delivery system and that implementation experience dictated changes to the specification in order to improve future implementations (assuming that the spec failed to specify the correct parameters). Greylisting is far more obviously a protocol with testable interoperability than is, say, RFC 5234 or 5322.

ESMTP is the protocol. Greylisting itself has no protocol (it's just a local policy for a mailserver using the ESMTP protocol) so there's no way to test interoperability.

You can test ESMTP interoperability, you can certainly test interoperabilty between ESMTP peers that are implementing greylisting. But you can't test *greylisting* interoperability, as there's no protocol there.

Cheers,
  Steve