If you use dmarc - ok, but spf by itself is a blunt instrument for the purpose you describe
~all was designed to account for the specific use case where most mail originates from the specified sources but a percentage may come from elsewhere, and that percentage is not worth tracking down or not easy to track down. Unless you're a large or even small to midsize sender or receiver (personal domains, small businesses running on their own mailserver, rare though that breed has become) you just might not care where else your mail comes from, or need the detailed metrics a dmarc implementation gives you. --srs > On 08-Jan-2016, at 12:17 PM, Mark Foster <blak...@blakjak.net> wrote: > > Surely deploying ~all is done in order for you to determine which edge cases > won't pass, prior to considering a move to -all? > > If it isn't -all then whilst it can be used for scoring it cannot alone be > the factor that determines accept-or-not. Beyond that if you choose to reject > email from my ~all-marked domain _because_ of ~all you are not compliant with > the RFC. > > SPF adherance is about preventing domain impersonation, more than about spam > filtering. What happens should ultimately be based on the spf record > published and not the receivers choice (or the receiver takes their chances > with collateral damage just like any other measure they might choose to take > on their platform). > > >> On 8 January 2016 7:06:54 PM NZDT, Suresh Ramasubramanian >> <ops.li...@gmail.com> wrote: >> That'd be wise, actually. SPF has seen some wildly inconsistent >> implementations of edge cases, from one domain to another. And that's not >> changed all that much. >> >> --srs >> >>> On 08-Jan-2016, at 11:19 AM, Simon Lyall <si...@darkmere.gen.nz> wrote: >>> >>> Soounds like unless I'm 100% sure every email I send will match the SPF I >>> should remove the record? >> >> >> mailop mailing list >> mailop@mailop.org >> https://chilli.nosignal.org/cgi-bin/mailman/listinfo/mailop > > -- > Sent from my Android device with K-9 Mail. Please excuse my brevity.
_______________________________________________ mailop mailing list mailop@mailop.org https://chilli.nosignal.org/cgi-bin/mailman/listinfo/mailop
