[MlMt] Bounce emails?

[Bill Cole]

On 8 Jun 2012, at 20:49, Joachim Tingvold wrote:

> On 9 Jun 2012, at 2:37, Bill Cole wrote:
>> That sort of feature in a MUA is fundamentally misguided. As 
>> described it is an exercise in intentional deception.
>
> Sure. Abusing completely legit features can be said about a lot of 
> things (-:
>
> That doesn't mean we have to sweep it under a carpet, using "it can be 
> abused" as the excuse for it.

I guess I wasn't clear...

What you describe is an intrinsically abusive feature. Such a "feature" 
has been offered in Eudora, Pine, Mutt, Claris Emailer, and Mail.app but 
that fact does not change the fact that it is an essentially wrong thing 
to do. A MUA that injects a message into the transport system should 
always do so with clear identification of itself.

A header-adding resend (e.g. RFC2822/5322) is significantly less 
problematic, and that was defined in RFC2822 because of the real-world 
experience with MUA's doing the simplistic re-injection of unmodified 
messages into the transport system with new arbitrary envelope 
recipients. In the modern world, there is also a practical problem 
because some spam filtering systems have figured out ways to do sanity 
checks on Received headers looking for forgeries, and will see the 
discrepancies of a pristine re-injection.