From: ext Dave CROCKER [mailto:dhc(_at_)dcrocker(_dot_)net]
Sent: 11 March, 2010 01:59
To: Polk, William T.
Cc: DKIM IETF WG; Eronen Pasi (Nokia-NRC/Helsinki); IESG; draft-ietf-
dkim-chairs(_at_)tools(_dot_)ietf(_dot_)org; Sean Turner
Subject: Re: An alternative way forward for dkim-deployment...
On 3/10/2010 3:18 PM, Polk, William T. wrote:
First, I do think it is harmful when document content does not match
(explicit or implied) scope.
And since I quoted text that made clear that the scope was constrained,
Again the question: just how much bullet-proofing against careless
I certainly had different expectations
based on the title and introductory material. In retrospect, I should
have brought my discuss up another level of abstraction but I
got focused on the details (what was missing) instead of why it
And yet even your original and later notes detail lacked detail, since
no way of knowing exactly what problem you were trying to address /or/
would satisfy and you did not provide clarification.
Personally, I would prefer to see this document expanded and either
address the various topics that are covered elsewhere or explicitly
reference the various bits of the DKIM overview and protocol specs
the rest of this information resides.
Since the document is loaded with citations, including the other DKIM
once again I have no idea what you mean.
Is there a chance that I am carrying my concern for the reader too
Not only "too far". As two of my earlier responses suggested,
additional text in line in response to the concern you initially stated
downside, given the intended audience. It would have been helpful for
address that point, over the last 5 weeks.
Perhaps... If you assume that the reader has already consumed the
protocol specs, then I am way off base. It certainly is not
that the wg members weren't confused. I just don't make that
Again, the text I quoted you earlier today is rather pointed in
role of this document as being to augment, rather than replace, other
My concern for the reader is based on my perspective at NIST.
guidance is actually something we do here, and sometimes we may even
The IETF has some history here, too. RFCs with the word "deployment"
date back to 1992. Interestingly, the first was about X.500 and X.400.
In any event, your statement here suggests that you are relying on the
culture for your concern, rather than the IETF culture.
However, they might use google to search on dkim and deployment
and try to use this document.
People "might" choose to do an infinite number of unreasonable things.
document out of context is only one of them. We cannot protect against
unreasonable action by a reader.
And, again, the document states its nature and goals explicitly.
I don't like holding this document up, especially when there are a
variety of simple solutions.
To say "solution" is to take as a given that there is a problem. You
responded to the responses that critique that assumption.
is clear this document will not evolve into "some sort of standalone,
complete instructional missive" so let's be clear about the document
scope, and move on.
The document /is/ already clear Tim, per the text that I quoted. It
needs the reader to be paying attention.