On 8/30/2013 2:37 PM, Hector Santos wrote:
> On 8/30/2013 10:46 AM, Tony Hansen wrote:
>>
>> The document describes a model for reputation services, particularly
>> those being produced by the Repute WG. It follows the recommendations
>> of RFc4101 for describing a protocol model, which requires answers to
>> 1) the problem the protocol is trying to achieve, 2) the meaning of
>> messages transmitted, and 3) important unobvious features of the
>> protocol. This document accomplishes its goals quite well.
>
> As a high potential implementator of this DKIM-REPUTE framework and
> user of any future REPUTE products based on this DKIM-REPUTE
> framework, I am somewhat disappointed to find not a single integration
> consideration for the highly adopted SPF technology. Not a single
> mentioning of the word or the integrated use of this highly adopted
> mail transport augmented technology. ...
Hector, what you're suggesting would be fine for a document that
specifically was written about reputation servers for email services.
But draft-ietf-repute-model is not the place for it.
While draft-ietf-repute-model does contain an example of how a
reputation service could be used to help assess a DKIM identifier, that
is not the thrust of this document.
RFC 4101 (Writing Protocol Models) contains this advice:
3.2 Abstraction is good
... try to abstract away pieces ...
3.3 A few well-chosen details sometimes help
... it's often a good approach to talk about the material in the
abstract and
then provide a concrete description of one specific piece to bring
it into focus. ...
The DKIM example is just that, an example of one way the model could be
used. It is strictly illustrative and not exclusionary in any way.
Tony Hansen
