ietf-dkim
[Top] [All Lists]

Re: [ietf-dkim] Last Call: draft-ietf-dkim-deployment (DomainKeys Identified Mail (DKIM) Development, Deployment and Operations) to Informational RFC

2009-12-04 14:24:30
The IESG wrote:
The IESG has received a request from the Domain Keys Identified Mail WG 
(dkim) to consider the following document:

- 'DomainKeys Identified Mail (DKIM) Development, Deployment and 
   Operations '
   <draft-ietf-dkim-deployment-09.txt> as an Informational RFC

I have some minor comments and nits, as follows:

* In sect. 2.4, the last but one paragraph discusses the relation 
between the RFC5322.From domain part and the d= signature tag. Perhaps 
a forward reference to Section 7.3 may be helpful.

* In sect. 2.3, a paragraph says

  The challenge in using additional layers of sub-domains is whether
  the extra granularity will be useful for the assessor.  In fact,
  potentially excessive levels invite ambiguity: if the assessor does
  not take advantage of the added granularity, then what granularity
  will it use?  That ambiguity would move the use of DKIM back to the
  realm of heuristics, rather than the deterministic processing that is
  its goal.

The rhetorical question it puts is not clear until one reads the 
following sect. 4.3. I would suggest to replace it with a plainer 
explanation, such as "because assessors might unilaterally decide to 
only use some rightmost part of the identifier."

* I'm unable to understand sect. 3.4 "Third Party Signer Key 
Management and Selector Administration". It probably means that 
messages are addressed to "prospects(_at_)companyA(_dot_)example" so that the 
MSA 
skips signing them, and signatures will be added by the exploding MTA. 
However, it might also mean that messages are already signed on 
submission.

* Sect 3.5, 1st par, s/a public key pair/an asymmetric key pair/.

* In sect. 3.5.2, bullet 4.3, I'd s/empty "p=" field/empty public-key 
data ("p=") field/.

* Sect. 4.2 says "The signing module uses the appropriate private key 
to create one or more signatures." Adding "(see Section 6.5)" may 
help; however, sect. 6.5 doesn't specify when signatures can be 
produced using the same key. Or did that mean a DomainKey-Signature?

* Also in sect. 4.2, in "key information SHOULD NOT be hard-coded into 
software", I've been doltishly wondering whether a hard coded path 
name such as /etc/dkim.key would be considered "key information". "Key 
data" (the numeric value of a key) should be less ambiguous, or are 
those terms used interchangeably?

* Sect. 5.4 references "external reputation or accreditation services" 
without giving any examples. Isn't it worth to mention RFC5518?

* Sect. 6.2 
s/i=marketing.domain.example/i=(_at_)marketing(_dot_)domain(_dot_)example/.

* Also in sect. 6.2, perhaps it is worth to note that parent domain 
signatures are not considered author signatures, e.g. adding "See also 
Section 7.3.2".

* In sect. 6.4, the last but one paragraph discusses third party 
signatures, but it misses an example. Mentioning that the signature 
would use "s=benefits; d=companyA.example" would make it more easily 
comparable with the paragraph that follows.

* Sect. 6.5 doesn't mention changing algorithms, as described in A.2, A.3.

* Sections 7.3.2 and 7.3.3 contain very similar text:

                          For example, if the address in the
  RFC5322.From is bob(_at_)company(_dot_)example, the SDID value of the author
  signature must be company.example.

and

  For example, email with an RFC5322.From address of bob@
  companyA.example MUST have an author signature where the SDID value
  is "companyA.example" or it will fail an ADSP validation.

In addition, that example is not overly explicative. A variation might 
say, e.g.
                                    For example, if the address in the
  RFC5322.From is bob(_at_)subdomain(_dot_)company(_dot_)example, the SDID 
value of the
  author signature must be subdomain.company.example, not the parent
  domain.

* (In sect. 7.3.3 "Appendix A" and "Appendix B" in the last but one 
paragraph get the wrong HTML link, internal to the same document. Can 
that be avoided?)

* Sect. 7.5 s/i=clienta.espa.example/i=(_at_)clienta(_dot_)espa(_dot_)example/.

* In sect. 8.1, the paragraph quoted below mentions "multiple address" 
as a feature of both ISPs and mail clients. Probably it is enough to 
s/mail clients/ISPs/, since mail client features are not relevant 
here. However, an ISP may just allow any RFC5322.From address (omit 
checking) or sign messages according to the address they use (possibly 
featuring 3rd party signing), and the point that distinguishes ISP-1 
from "some other ISPs" is not made.

  For example, assume Joe works for Company A and has an email address
  joe(_at_)companya(_dot_)example(_dot_)  Joe also has an ISP-1 account
  joe(_at_)isp1(_dot_)example(_dot_)com, and he uses ISP-1's multiple address 
feature to
  attach his work email joe(_at_)companya(_dot_)example to his ISP-1 account.
  When Joe sends email from his ISP-1 account and uses
  joe(_at_)companya(_dot_)example as his designated RFC5322.From address, that
  email cannot have a signature with d=companya.example because the
  ISP-1 servers have no access to Company A's private key.  In ISP-1's
  case it will have an ISP-1 signature, but for some other mail clients
  offering the same multiple address feature there may be no signature
  at all on the message.

_______________________________________________
NOTE WELL: This list operates according to 
http://mipassoc.org/dkim/ietf-list-rules.html

<Prev in Thread] Current Thread [Next in Thread>
  • Re: [ietf-dkim] Last Call: draft-ietf-dkim-deployment (DomainKeys Identified Mail (DKIM) Development, Deployment and Operations) to Informational RFC, Alessandro Vesely <=