[ietf-dkim] ISSUE: dkim-overview -- normative statements

[Douglas Otis]

On Jul 16, 2007, at 3:03 AM, Charles Lindsey wrote:

> On Sat, 14 Jul 2007 21:01:34 +0100, Paul Hoffman  
> <paul.hoffman at domain-assurance.org> wrote:
>
>> Many thanks to Dave for bringing this up.
>>
>> At 2:55 PM -0400 7/14/07, Dave Crocker wrote:
>>
>> I think it would be fine to make this a standards-track document  
>> with normative language.
>
> I disagree. The purpose of an "Overview" is to give an informal  
> summary of the effect, and especially the reasoning and motivation  
> for, a collection of standards. So it may well outline what the  
> standards do, and indicate the sort of normative provisions they  
> would make (which can include indication of whether such provisions  
> are MUST/SHOULD/MAY).
>
> But it should include somewhere wording such as:
>
>      "This document provides an overview of the DKIM collection of  
> related standards, indicating how they are intended to work  
> together to produce [list of desirable effects]. But this document  
> is not normative itself; {RFCxxx], [RFCyyy] and [RFCzzz] should be  
> consulted for the detailed normative requirements."

An overview should not be normative.  Agreed.

At some point, corrections to RFCs describing the protocol might be  
required, or a BCP document to explain desired methods to address  
various issues might be appropriate, but much later.  This overview  
has not been reviewed from a perspective of noting whether it creates  
new or conflicting normative requirements.  Expectations were just  
the opposite.

As a side note, would it be possible to reference the TPA-SSP instead  
of the DOSP draft in the working drafts section on the DKIM website?   
TPA-SSP will be easier to understand.  You should have the XML and  
HTML versions of this draft.

-Doug