william(at)elan.net wrote:
How about this, if somebody wants to start different scope
that does not include records from first one then HAVE TO
close existing scope
That's another idea. Roger's "a new sc= implicitly closes
any old sc=" is probably simpler.
"v=spf1 sc=s1 m1 sc=-s1,s2 m2 m3 -all"
That's the same as "sc=s1 m1 sc=s2 m2 m3" in Roger's notation.
Messy if there are many scopes and the order of mechanisms is
significant, if you want say m3 in all (out of three) scopes:
"v=spf2 sc=s1 m1 sc=s2 m2 sc=s1,s2,s3 m3 -all" (Roger)
"v=spf2 sc=s1 m1 sc=-s1,s2 m2 sc=s1,s3 m3 -all" (you)
"v=spf2 m1 sc=s1 m2 sc=s2 m3 -all" (protocol-03)
Mark's notation has a tendency to be the shortest, but not the
best readable. Maybe Mark had a HP pocket calculator like me,
and we're used to weird stuff like reverse Polish notation. ;-)
"v=spf1 sc=s1 m1 sc=s2 m2 m3 -all"
"v=spf2 sc=s1 m1 sc=s1,s2 m2 m3 -all" (Roger)
"v=spf2 m1 sc=s1 m2 m3 -all" (protocol-03 sryle)
But your resp. Roger's idea works best with scopes, not with
"modify only one mechanism". The protocol-03 / Sender-ID style
is shorter and more general, but less readable. Allowing both
styles could be really confusing. But that's a spf2.0 problem.
For a PRA option or your "eh" in v=spf1 we need IMHO op= (p=).
Bye, Frank