Re: [ietf-smtp] DSNs

[Dave Crocker <[e-mail]>]

On 4/25/2020 11:47 AM, Viktor Dukhovni wrote:
> I'm very much with Dave on this one.  If you're not offering
> NOTIFY=SUCCESS, then you're not offering DSN.  Pretending to support it,
> but then not offering it is breakage.  On the other hand, since I ignore
> "DSN" on the outbound leg, I avoid running into any problems should the
> remote side misrepresent its behaviour.


To be clear, I wasn't commenting on the details of what should be in the 
core, in this particular case.

But since you've raised it, I'll note that a core which has 
well-established failures to gain widespread adoption and use is a 
pretty poor core.

A specifications is not an exercisesin dictating ideals but in 
facilitating practice.  When there is established practice, it needs to 
return the favor by dictating the specification.

d/

-- 
Dave Crocker
Brandenburg InternetWorking
bbiw.net