spf-discuss
[Top] [All Lists]

Re: [spf-discuss] Per/user policies in "Large Domains" (was Fixing Forwarding with RPF)

2006-11-15 07:41:14
On Wednesday 15 November 2006 13:52, Scott Kitterman wrote:
I think you miss a key point....

SPF is designed to work at the domain level for domain owners.  It offers
the exists: mechanism for more finely grained policies when necessary, but
this requires a custom DNS back end that is complex to implement correctly.
 Pobox.com had a beta of this, but no one, AFAIK, has gotten this into
production.  This is a hard engineering problem (I'll be glad to say I was
wrong when someone shows running code).

There are two "domain owners" here. There is a small domain being forwarded to 
a "large domain" (as used in RFC4408 9.3.3). I am trying to make it friendly 
to both types of domain. There may well be a complicated solution, but a 
simpler solution should be better.

IIRC, we were talking about cheap providers.  Cheap providers aren't going
to do this.

Probably not.
From reading your last message it isn't clear to me that you understand the
difference between REJECT and BOUNCE.  This is essential.

As I understand it, a REJECT happens when an MTA refuses to accept the 
message. A bounce happens when the previous MTA attempts to advise the sender 
of non-delivery. Is that right?

10 years ago open relays were the norm.
5 years ago bouncing was standard.

Today bouncing is considered anti-social at best.  Forwarding to much spam
will get a forwarder blacklisted.

Depends who is doing the considering. Some people on this thread see bouncing 
as standards-compliant, but whichever way, the question is whether turning 
SPF off on a particular mailbox could generate one. I don't see how it could, 
though, as has been pointed out, if someone did that and then put an 
autoresponder on the box, that could have a similar effect. So it's bad 
practice to use autoresponders unwisely. Is that an SPF issue? If so, SPF 
should lay down good practice.

My forwarder is just doing its job, unfortunately, when it forwards spam 
addressed to me. I have never known it to originate spam, which is what ought 
to get it blacklisted. To get it blacklisted on a respectable list someone 
has to have a legitimate complaint, and complaining about a service I have 
set up for myself doesn't strike me as legitimate. In any case, what has this 
to do with SPF? We know SPF is no use on a forwarded message.

Even if your analysis was correct and your proposal would not result in
significantly more bounces, the trends are against you.

What trends? 
No bounces except to a verified mail from is the best practice today.

I agree.
Any new proposal really has to have essentially zero backscatter risk.

I think it has. That's what I'm asking my detractors to demonstrate is wrong 
with my suggestion. So far, they have not.

At the end of the day, my incoming mail is my concern. SPF is about senders, 
and should not be forced upon receivers when they don't want it. (Or, more 
accurately, where they don't want it. I would be delighted if my forwarder 
adopted it, and maybe they will one day.)


KJP

-------
Sender Policy Framework: http://www.openspf.org/
Archives at http://archives.listbox.com/spf-discuss/current/
To unsubscribe, change your address, or temporarily deactivate your
subscription, please go to http://v2.listbox.com/member/?list_id=735

-------
Sender Policy Framework: http://www.openspf.org/
Archives at http://archives.listbox.com/spf-discuss/current/
To unsubscribe, change your address, or temporarily deactivate your 
subscription, 
please go to http://v2.listbox.com/member/?list_id=735

<Prev in Thread] Current Thread [Next in Thread>