Some comments on 
https://datatracker.ietf.org/doc/draft-ietf-marf-as/?include_text=1

Most of the document looks fine, but I have some concerns about the sections 
that deal with unsolicited reports.

Section 5. Solicited and Unsolicited Reports

"Scored high by spam filters" really isn't a good reason to send an abuse 
report. Sending mail that "looks like" spam according to a mechanical content 
filter isn't an AUP violation, and reports about it aren't going to be 
actionable as there's nothing wrong with sending mail with a high spamassassin 
score. While, arguably, reports of mail that "looks like" spam might be a basis 
for investigation by an abuse desk as to whether a customer is doing something 
wrong it's not a very useful basis - if there is a source of mail on your 
network that's found offensive by recipients, you'll find out fairly quickly by 
receiving informed abuse reports, manually or from this-is-spam-button driven 
feedback loops, so reports of mail that "looks like" spam aren't going to 
provide much data that the abuse desk won't already have via other channels. 

Non-actionable reports - whether they be false positives or not - reduce the 
efficiency of an abuse desk. Let's just not mention reports of mail scored high 
by spam filters, and quietly leave it as part of "anything else the sender 
considers".

(Automated reports of spam-like content may well be useful as part of an "avoid 
false positives" sort of feedback loop, but not as an abuse report. And 
automated reports of mail that has virus or malware content might be 
appropriate, even though spam-like content isn't.)


The DKIM signer of a message is definitely a good target for solicited reports 
as part of a feedback loop. For unsolicited reports it's much less clear that 
that's the case. There are some mechanical reasons - the d= value is an 
identifier for reputation, not a domain intended to receive email, so we don't 
want to imply that [email protected] is automatically a good 
place to report mail that's signed with d=transactional.blighty.com. In a 
broader sense, though, it's likely that the DKIM signing domain is the author 
of the mail - and if the mail is objectionable, they're often the least useful 
place to notify. In the case of mail sent through an ESP, the DKIM d= domain is 
likely to end up at the customer, rather than at the ESP where it's more likely 
to be handled correctly.

That's one instance of a more general issue of sending automated, unsolicited 
reports. Identifying an appropriate recipient for a report is *hard* to do 
automatically, while "shotgunning" reports to multiple email addresses because 
one of them might be appropriate is unhelpful. Sending them in a format that's 
intended primarily to be machine-readable and handled automatically by the 
entity receiving the reports leads to the problem of knowing what email address 
at what entity is intended to handle machine-readable reports. There's no 
standard for it, and there's a risk that those machine-readable reports will be 
discarded or ignored if they're sent to role addresses chosen at random. 

Section 8 discusses this issue again, and somewhat better. Would it make sense 
to move the mention of DKIM to section 8, where it would be in a better context?





Section 7. Receiving and Processing Complaint-Based Solicited Reports

"Point 3 - Mail operators SHOULD utilize an automated system to receive and 
process these reports". 

That's really an engineering and operational decision. If I only receive two or 
three ARF reports a day, and my MUA or ticketing system can display them 
correctly, then handling them as part of my usual customer support / abuse desk 
workflow is likely to be more reliable (and much less expensive to implement) 
than an automated system. An ARF based FBL is still of value to me, but 
automating it would be a poor decision. RFC 6449 Section 4.4 says pretty much 
that. I think we'd be better just pointing people at that section of RFC 6449, 
rather than saying that automation is a SHOULD.





Section 8. Generating and Handling Unsolicited Reports

Is authentication failure abusive? If it is, we should say why (and in what 
context). If not, it shouldn't be in this document.





I'm not sure it's true that "even recipient systems that haven't asked for ARF 
format reports handle them at least as well as any other format such as plain 
text, with or without a copy of the message attached."

I know for sure that there are ticketing systems in use that don't handle 
anything other than plain text particularly well. They will not handle ARF 
reports as well as they would handle plain text. 

Unexpected multipart/* formats (such as the multipart/report type used by ARF) 
aren't necessarily going to be handled well even by MIME aware MUAs. Similarly, 
message/feedback-report isn't necessarily going to be handled well by MIME 
aware MUAs.

As an example, here's a snapshot of an ARF report displayed by Mail.app: 
https://skitch.com/steveatkins/g2bkt/inbox-27642-messages-33-unread

And here's one displayed by gmail: 
https://skitch.com/steveatkins/g2bck/inbox-27642-messages-33-unread

And Outlook Express: https://skitch.com/steveatkins/g2bpj/192.168.80.89

All three of those are more sophisticated than the web-based ticketing systems 
commonly used at smaller ISPs.

(Alpine, OTOH, displayed it just fine - 
https://skitch.com/steveatkins/g2bq8/steve-infrastructure-mail )

Ticketing systems are one of the last bits of the email system that don't 
handle MIME particularly robustly, if at all, and abuse desks are (quite 
reasonably) wary of clicking on attachments - all attachments, including 
attached message/rfc822. Quite a few ISPs (including national US ISPs) ask 
reporters explicitly to not forward email as an attachment, rather to paste the 
headers and body of the mail being reported into a new message, due to either 
tool limitations or a policy not to open attachments for security reasons. 

Everybody can handle the lowest common denominator of text/plain, but that's 
about it.

So… I don't think it's valid to say that recipient systems that are not 
advertised to accept ARF will render ARF format messages in a way that will 
lead to them being handled. At the very least, anyone sending unsolicited 
messages in ARF format must assume that recipients will not be able to see the 
ARF metadata, and should include all information needed in human readable 
format in the first, text/plain section of the report. Also, they should ensure 
that the report is readable when viewed as plain text, to give low end 
ticketing systems as much help as possible - that might includes advice on 
appropriate MIME encoding and choice of character sets and so on. And they 
should be aware that the report may be discarded or ignored due to it's format.

I don't think that experience purely of sending unsolicited ARF reports can 
really lead to the claim that all recipients handle them well - as the 
recipients that *don't* handle them well are simply going to discard or ignore 
the report, rather than trying to work out why it was sent. Abuse desks 
generally have more than enough actionable reports in their queue to waste time 
trying to decipher ones that they can't read.

IIRC, John includes a link to a web-based rendering of each report in the plain 
text part of the unsolicited reports he sends out (not a bad idea, if you have 
the infrastructure for it), so they may be more widely readable by abuse desks 
than an ARF format report would be.

Cheers,
  Steve

_______________________________________________
marf mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/marf

Reply via email to