[feedback-report] Let's get started

[Murray S. Kucherawy]

> -----Original Message-----
> From: Yakov Shafranovich [mailto:yakov at shaftek.org]
> Sent: Monday, December 28, 2009 6:12 AM
> To: Murray S. Kucherawy
> Cc: abuse-feedback-report at mipassoc.org
> Subject: Re: [feedback-report] Let's get started
> 
> [...]
> 1. Were we planning on putting the Domain Keys stuff into a companion
> draft instead of the main document? Then the following from section
> 1.1 should be removed:
> 
> "To advise providers that certify or otherwise make assertions about
> mail of recipient disagreement with the assertions."

Yep, it was a reference to the now-removed DKIM support.  Removed this text.

> 2. Section 1.1:
> "To inform email service providers or other primarily outbound
> 	senders that there may be issues regarding their mail; these
> 	issues include (but are not limited to) reports that the mail may
> 	be considered to be "spam" by a recipient of the message;
> "
> This may be too generic, perhaps we should narrow this specifically to
> abuse and related issues?

Fair enough.  I've inserted the word "abuse" before the first "issues", and that seems to cover it.

> 3. Section 2.f.:
> "which the report is being generated and MAY include only a typical
> forwarding prefix"
> 
> Is there an existing standard we need to reference that covers
> forwarding?

I don't know of one, which is in fact why I changed the word "standard" to "typical".  Anyone else know of one?

> 4. Section 4 - spelling error in "report types":
> 
> "Implementations using new rpeort types"

Fixed.

> 5. Section 6.6 - redaction:
> 
> Do we need a new header that indicates to the report recipient that
> the message was redacted? There is a header "Incomplete-Copy" in RFC
> 4021 which has a similar purpose for X400/email mappings.

I'd be fine with including such in the second MIME part that contains ARF-specific header fields.  Anyone else have a thought on this?

The specific header field you give is defined in RFC2156 but unfortunately there's no description of what specifically it covers.  RFC4021 says it has to do with an incomplete body though, which is generally not the case here.

> 6. A general question that was once raised on the mailing list - do we
> need to take into account RFC 3864 and the message header field
> registries?

I don't think so because we're not registering header fields that are used in header blocks of messages.  The header fields we're defining are part of the message body (a MIME part, in particular).  DSN (RFC3464), for example, did no header field registrations; all of the header fields it created are defined there, and it even created its own registries for values of a few of them.  I think that's a pretty good analogy for what we're doing with ARF.

Thanks for the review!

-MSK