With tongue in cheek, lacking all ETOH, and just a wee cafiene, spake I.
Where I am from that which appears to be a duck in every way is definately a
duck, identified positively as such, and redmond is a rather wet place.
In the beginning, there were three guys in a college dorm...
Later... There was but one...
Solid embrace 'tween he and IBM in the form of cooperative development to the
form of OS/2.... But wait. We must now test, debug, rinse & repeat says
one. Nay says he, we must bundle and sell, thus be paid for the privilege
of debugging our creature, and be paid again for the fixes... Nay, says IBM,
and he ran home to release his bugs and be paid for it.
Upon Preparatory release he applied to the U.S. Patents office and was denied
as 'Windows' was a common term, already in use for (come to think of it, most
of you, as well as my Great Grand Daddy glazed a pain or two!), who knows
exactly how long.
Release as Windows any way, sell, get paid, fix bugs, introduce more, more
pay, fattening the cow... ( ten generations later they are still rinsing &
repeating the same old methods).
Then comes 'Partnerships' and 'Agreements', etc... That are again broken...
Task forces joined. When a full out adoption is not agreed to, they once
again run, release to the masses, thus gaining 'de facto standard' status...
As it is used enmasse albiet on the M$ platform...
(Sun java is just another reminder...)
(H'MM, there seems to be some audio, and some browser thingies here as
well...)
XML has it's time and place, which I do not believe is anywhere near DNS.
(This is my personal opinion, if you need XML, it needs a private, separate,
and distinct carrier (another show of force by M$' not-invented here
mentality)).
Ad in finum...
And let it be said that, "Smitty said, soon it will come to pass that Meng
will become just another notch on the M$ Cannon..."
Now, If you have read this far, I'll drop into a serios tome...
Several weeks ago I sent an off-list email to Meng with no response.
Here is the meat of that mail...
BTW: We have already implimented SPF, long before the M$ debacle...
1) All current MTA's have the ability to place custom headers in a mail pack.
2) All current MTA's have a method of calculating a checksum/hash, based on
set criterion, and using same in #1.
3) All current MTA's have logging capabilities, as well as custom log
formats.
4) All ISP's/Commercial Entities (read as legitimate), already have logging
and query/response capabilities (they may call them by other names...)
5) This would require no major modification to the SPF spec, a simple
addition of one additional field in the txt response...
"v=spf1 mx -all DottedQuad:Port"...
and an agreed upon header in the mail pac, and a minimum on-line time of the
HASH/RECIPIENT query data.
Here is the beef!
a) Responsible party for each MTA publishes their SPF txt record as #5 above.
b) Party 'A' (AKA: SENDER) sends a mail to Party 'D' (AKA: RECIPIENT).
c) MTA 'B' (AKA: ORIGINATING MTA responsible for Party 'A') validates the
user as one of theirs, calculates the chechsum (based on the sender,
recipient, and a locally known secret(or secrets)) and places the appropriate
header in the mail pack, then forwards the mail on to the appropriate
destination MTA.
d) MTA 'C' recieves the mail, validates SPF and aquires the txt record (as in
#5 above). No further DNS lookups are required (that is why IP, NOT FQDN).
Request is sent to MTA 'A' at IP:port with the payload being HASH (from
header) and RECIPIENT, and recieves back a response of Pass or Fail (1/0,
T/F, whatever is agreed upon in the spec).
e) Responsible party for MTA 'C' decides what to do with the mailpack...
I will post this to the dev list as well, maybe it needs some heavy kicking
around, then again, maybe I do!
Sincerely & Be Safe!!
Smitty