On 12/Sep/11 16:28, John Levine wrote: >> Consider >> _report.myunknowndomain.name. TXT "[email protected]" > > Speaking both in general, and specifically as Mr. abuse.net, this is > precisely what I want to forbid. I cannot tell you how many lazy or > incompetent ISPs have foisted off some problem by telling their users > to report it to abuse.net, which is of course completely useless and > a waste of both the user's time and mine.
OTOH, such attitude proves that there is a market for that. According to plan "A", users should send complaints to their Mailbox Providers. It is the MP's abuse team who may opt for forwarding to abuse.net: they should be smart enough to understand a clause such as "do so only for domains that (a) are authenticated by dkim or spf, and (b) explicitly authorized by abuse.net". Possibly, abuse.net has a contract with such domains, e.g. providing for higher prices according to the number of complaints. NB: I'm not implying Mr. abuse.net has to be happy of doing such business. It is just a possibility that you and/or someone else might put to work. > I realize that we could invent some DNS thing with extra text records > that are supposed to cross reference in some way, but I see no reason > to invent yet more mechanism. If you've outsourced your abuse > handling to someone else, we already know how to add a single line in > your mail config to forward your abuse address to them. At last that > way when some bozo forwards his complaints to me without asking, I > only have one address to blacklist. The point is that the feedback generator has to be aware of that forwarding, in order to trust sending reports there. As an alternative to inventing more mechanisms, we could tweak VBR. Assume myunknowndomain is vouched for by abuse.net. Then one'd expect abuse.net to accept complaints for abuse originating there, but to what local part should they be addressed to? The reporting-discovery could say that an external domain is acceptable if and only if the latter vouches for myunknowndomain. Or it could modify VBR so as to allow, e.g., "all r=bofh". These two are more or less equivalent variations, even if they may seem to avoid inventing one more mechanism. _______________________________________________ marf mailing list [email protected] https://www.ietf.org/mailman/listinfo/marf
