On Fri, Mar 10, 2006 at 11:20:56PM -0500, Scott Kitterman wrote:
I currently would opt for binary only but I may easely change my mind.
The SPF RFC already defines TYPE99 as a TXT format record, so if you want a
binary record, then you would have to get another new RR Type, wouldn't you?
No.
I see no problem. The current RFC defines the content for type99 but
only for v=spf1.
Type99 binary would not be recognized by v=spf1 parsers, thus discarded
similarly as "hello" is discarded as possible spf record.
v=spf2.1 parsers will recognize binary format.
v=spf2.1 parsers will also be able to recognize type99 text-like records.
No clash here.
But I agree, TYPE99 only for the new version would be the goal. I'd still
allow for dual publication in TYPE99 and TXT because non-BIND resolvers are
not all going to support TYPE99 for some time, but I don't think we should
require it.
Allow dual publication again and we have the same s**t as we have now.
Either allow it and live with it forever, or do not allow it. You can't
allow it but aim for the opposite.
Worst case is to allow but not require dual publication:
a) publishes using TXT records only
b) publishes using SPF records only
c) publishes both
A) only looks for TXT records
B) only looks for SPF records
C) looks for both types
(A) has such an old, disfunctional, resolver. (A) cannot look for (b)'s
records.
Thus, (b) cannot exist. Most people will be (a), not (c). Much software will
be of type (C). Result: lots of wasted resources.
IMO we should make the SPF record type mandatory or we should
abandon it entirely (both for version>1 of course).
Alex
-------
Sender Policy Framework: http://www.openspf.org/
Archives at http://archives.listbox.com/spf-discuss/current/
To unsubscribe, change your address, or temporarily deactivate your
subscription,
please go to
http://v2.listbox.com/member/?listname=spf-discuss(_at_)v2(_dot_)listbox(_dot_)com