[ietf-dkim] Not exactly not a threat analysis

[Keith Moore]

>>>>And because "abuse" is subjective (one recipient's spam is another
>>>>recipient's useful ad), you end up both legitimizing some amount of
>>>>abuse and marginalizing useful and valid behavior.
>>>
>>>I don't see any clear signs of the convergence to mediocrity that you
>>>are concerned about,
>>
>>Well, to me the amount of mail being discarded due to false indications from
>>blacklists is a clear sign of mediocrity.
> 
> 
> I'd just say that that's a sign of a poorly implemented policy :-)

If you mean that trusting someone else to decide, on behalf of a large 
and diverse set of users, what is good for all of those users, is a 
poorly chosen policy - then I emphatically agree.

> The problem is that spam filtering is AI-complete and therefore cannot be
> implemented perfectly. This is why I'm trying to encourage people to avoid
> thinking in terms of the details of particular kinds of bad email. Looking
> back at your first sentence above I notice that you have mis-identified
> the abuse: it is not the fact that the message is an advertisement, it is
> the fact that it has been sent to someone who didn't want it. 

I don't see how I've mis-identified the abuse, as I've been consistently 
saying that spam is related to who _sent_ the message to the recipient. 
  (As opposed to phishing, which is unrelated to who sent the message).

Keith