On Sat, 2004-10-30 at 16:27, guy wrote:
Existing spf1 systems would not parse this.
Can you be more specific as to the actual problems you're referring to?
I'm proposing if a person wants to have "a mx" read in both mailfrom and
pra scope for people who implement both, that the spec would allow, and
it would recommend, the person publish following two TXT records:
"v=spf1 a mx -all"
"v=spf1/mfrom,pra a mx -all"
Alternatively, they could publish the following two records, still
within the same recommendations:
"v=spf1 a mx -all"
"v=spf1/pra a mx -all"
Current spf1 systems ignore each of the two second lines, but current
spf1 systems don't understand the pra scope anyway, (and my proposal
doesn't define it.)
So where's the incompatibility with what my suggestion recommends?
(If you're referring to the idea of multiple SPF-type RR's as the source
of incompatibility, well, we don't have an SPF RR yet anyway, much less
multiple ones, so implementations haven't yet had a chance to actually
develop incompatibilities. But even if we did already have a single SPF
RR, my suggestions work similarly compatibly with one SPF RR just as
with the one current TXT record.)
--
Mark Shewmaker
mark(_at_)primefactor(_dot_)com