[ietf-dkim] How to reconcile passive vs active?

[Dave Crocker]

Steve Atkins wrote:
>> From my perspective, the number needn't be small at all.  Small
>> organizations with their own mail processing infrastructure can with
...
> Even when it decreases overall deliverability? That is to say, causes
> legitimate email to be treated as forgeries and, likely, discarded.
> 
> I can see cases where that's going to be an appropriate tradeoff, but I
> don't think they're as widespread as some people think.


We need to find a way to discuss the design choices so that we are not faced
with a bevy of personal prognostications about the likelihood of particular
outcomes.  Most, if not all, of us will be wrong.

Rather:

Do we see a clear and compelling benefit for a design choice being proposed?

Is it substantially better than some other choice being proposed?

Do we see an absence of significant detriments? (Complexity, scaling,
performance, reliability, etc.)

Do we see a clear and substantial base of users/organizations for the choice NOW?

Only then might we consider adding a guess about the size of the population from
which this sample of adopting users/organization is taken.

Debating whether my sample is bigger than yours is certain to be unproductive.

d/
-- 

  Dave Crocker
  Brandenburg InternetWorking
  bbiw.net