On Sat, May 21, 2005 at 08:43:40AM +0100, Chris Haynes wrote:
Actually, to me this is not trivial, not if we want things to be clear for
new adopters.
Agreed.
When I first came across SPF this notation really puzzled me. It took me
several readings to realize that these single characters, which looked
like they were minor operational prefixes or syntactic qualifiers, where
actually _the result_ of a match to the mechanism they preceed. It was
especially wierd because it is unusual to have the result as the first
part of a syntactic group - normally the consequence is to the right of
the cause (in LTR languages). That's why I think it is valuable to pick
some word that makes their semantic significance clear.
Something like '-a' _means_ if the IP address matches the 'a' mechanism
the _result_ of the attempt to determine the sender's SPF policy for this
IP is 'FAIL'.
This I don't agree with.
"FAIL" if "a" matches
"PASS" if "ip4" matches
"Do this" when "that".
One could argue (as I am now doing) that the symbol represents the
sender's _policy_ for any IP that matches the related mechanism. That
really makes its semantic significance clear.
So, I hereby amend my vote to 'policy' or 'result'.
On policy:
This little character does NOT present the sender's policy. It only
represents a (potentially small) part of it. Talk about confusion:
"Hey, but if that's the policy, why do I need more stuff in my record?"
Policy would be wrong.
On result:
"Do this" when "that"
could also be
"This happens" when "that".
I like "action" better than "result" but could live with result.
Alex