[Asrg] DNSBL BCP 04 Re: draft-irtf-asrg-bcp-blacklists draft updated.

[Chris Lewis]

I'm finally getting around to making the revisions for the DNSBL BCP 
draft revision 3, since there's a possibility that the BCP may be 
accepted at the upcoming Dublin IETF.

I'm going to try to get an 04 revision out today.

I'm working through all the NITs and minor wording changes, and 
accepting almost all of them as given. Eg: the list of changes from Tony 
Hansen.  Thanks all!).  _None_ of these change the BCP in any 
substantive fashion.

However, apparently, the IETF document site has frozen all submissions 
because of the impending IETF meeting, and won't unfreeze until Sunday 
or Monday.  Which means it won't be "officially visible" until then.

The only bit that generated a substantive discussion was Ian's mention 
of this:

> I'm not sure about this: "DNSBL providers SHOULD NOT be held
>    accountable in any way for the consequences of use of a DNSBL applied
>    in an un-intended way."
> 
> The implication is that providers may or even should be held accountable 
> for consequences of use of a DNSBL applied in an intended way.

My first reaction was "if I had meant that implication, I would have 
stated it explicitly".  But that's rather flippant.

There's an important point in that statement (eg: don't blame the PBL 
for people doing deep received header tracing), so we can't lose it. 
But subsequent wording suggestions didn't seem to quite fix the 
perceptual problem.

I've come up with a different approach - rather than trying to do a 
SHOULD/MUST etc, I'm just going to provide an existing example, worded 
thusly:

------------------------------------------------------
For example, one DNSBL requires, if the DNSBL is used contrary
to their usage instructions, that the DNSBL user should not identify
the DNSBL being used, and further that it is the DNSBL user's 
responsibility to mitigate the effect of the listing locally.
------------------------------------------------------

Whaddya think?

I could put a link into the specific DNSBL's policy that says that as a 
informative reference, but I've avoided calling out specific DNSBLs so 
far, and I'd like to keep it that way.