[ietf-dkim] ISSUE: Updating Section 6.5: Recommended Signature Content

[Murray S. Kucherawy]

> -----Original Message-----
> From: barryleiba.mailing.lists at gmail.com [mailto:barryleiba.mailing.lists at gmail.com] On Behalf Of Barry Leiba
> Sent: Tuesday, April 26, 2011 11:25 AM
> To: Murray S. Kucherawy
> Cc: ietf-dkim at mipassoc.org
> Subject: Re: [ietf-dkim] ISSUE: Updating Section 6.5: Recommended Signature Content
> 
> > It is redundant to the first sentence in the text of Section 6.4 which already
> > says From MUST be signed; there's no reason to repeat it here, especially
> > since this section is clearly advisory in nature only, and thus there's no
> > harm in removing it.
> 
> Perhaps, but I agree that it would be useful to remind people.  Why
> not leave it there, with a reference to the normative statement?

With the back-reference, I guess I feel better about it.  Without it, in my head it's akin to repeating normative assertions in different places, begging for divergence.

> > The issue is that adding things to DKIM now will prevent its progress toward
> > Draft Standard, which is something we're trying to avoid.  Removing things, on
> > the other hand, doesn't introduce any backward compatibility issues, and so
> > that's allowed.
> 
> I don't think so.  This section is not normative -- we've already
> *removed* the SHOULD in it, and that's the removal.  After that,
> anything we add is an idle suggestion, and doesn't affect conformance.
>  I think we can put anything we want in here, informatively.  We can
> even be general -- that is, we can say "Authentication-Results [tm]",
> or we can say, "any header field that conveys authentication results,"
> or some such.

If consensus concurs with that view, then I would indeed like to add something like "any header field that conveys important meta-information about the message, such as authentication information from a relay (e.g., [RFC5451])".