spf-discuss
[Top] [All Lists]

Re: Shortcuts to current mechanisms defined in RFC ?

2004-07-11 11:36:52
Andrew G. Tereschenko wrote:

What is a reason to consume valuable DNS answer space by
useless .0 (and .0.0 as well .0.0.0) values ?

Maybe robustness and readability.  If I'd ask Google for CIDR
the first result is...

<http://public.pacbell.net/dedicated/cidr.html>

| A CIDR address includes the standard 32-bit IP address and
| also information on how many bits are used for the network
| prefix

They name RfC 1517 .. 1519 as sources, and 1517 points to 1518,
and that points to 1338, and then I gave up to find some kind
of standard syntax (1338 still has the IP/netmask notation).

You ideas to simplify resp. compress the SPF syntax all make
sense for me, but maybe it would be better if you send your
ideas to MARID (mxcomp).

One idea discussed (or rather: posted) in the MARID group was
to use a binary format, where classic SPF and SederID XML are
used only locally as some kind of verbose representation resp.
interface to the binary format.

As far as I understood it they (= MARID) dropped this idea,
because a binary format with a new record type (the original
RMX idea) won't work in the near future.  For similar reasons
they also dropped the XML variant, because the classic SPF TXT
was good enough, and XML-over-DNS is a nightmare.

At the moment Weng and Mark work on the series of documents for
"united SPF" (as input for MARID), and therefore your ideas to
simplify the "text" format could go to MARID.  As far as I'm
concerned I could live with an improved v=spf2 in a final RfC.

But changing classic SPF only to save some bytes in a way not
compatible with existing v=spf1 implementatios would be a bad
idea.  Don't start optimizations before the real thing stands.

                   Bye, Frank
-- 



<Prev in Thread] Current Thread [Next in Thread>