In message
<236086(_dot_)1373244986(_at_)turing-police(_dot_)cc(_dot_)vt(_dot_)edu>,
Valdis(_dot_)Kletnieks(_at_)vt(_dot_)edu writes:
On Sun, 07 Jul 2013 23:17:56 +0100, Sabahattin Gucukoglu said:
RFC 2821 language said I was supposed to expect dotless domains (the text,
anyway) so I did. It's also consistent with the use of full email
addresses.
While I agree that advising people to depend on it would be a bad idea, it's
something that can be easily solved in code.
It's hardly as "easy" as you might think. Sure, fixing the next release of
MTAs to support dotless domains is maybe 15 minutes work per MTA.
The pain is:
1) Sites that haven't upgraded their MTAs yet. I have a fairly high-volume
Listserv box still running RHEL 4.0 running Sendmail 8.13.1 - and it will
keep running that until I finally get a chance to upgrade it sometime in
the next few months. And that's me. How long will it take for the sites
that *can't* spell Sendmail or Postfix to upgrade?
2) The ever-dreaded "local alias" problem. Until you've gone thorugh *every*
corporate and organization network and fixed every single user's address
book so it doesn't say 'fred@locus' but upgrade it to
'fred(_at_)locus(_dot_)example(_dot_)com',
and then fix their MUA to prevent the user from savin a local alias again,
you won't be able to really reliably use '@locus' as a dotless destination.
And the speed things get upgraded can be glacial. Non-SMTP examples follow:
1) IPv6 is now available at something like 60% of colos and a good chunk of
the 800 pound gorillas in the Alexa 1000. And yet, IPv6 traffic is still only
running 1-2% of total, simply because in a very high percentage of cases,
it won't work until the user finally replaces their CPE wireless router.
2) How many boxes out there are *still* on XP?
3) I recently did some monitoring of an IP address we used to have an NTP
server on. Over the course of an hour or so, I collected well over 300
unique IP addresses that still thought there was an NTP server at that
address. The interesting thing is that the hardware was shut down in 2002
(and the NTP server itself quite possibly in 2001, I'd have to check). So
literally a dozen years later, there were *still* hundreds of boxes in service
that had 12 year old configs.
So I'm pretty convinced it's not at all "just a small matter of code". At
best, it's "a small matter of code now to make it perhaps workable in 2025
or so".
And even after you fix the MTAs and the MUAs you also have to stop
"ssh/telnet/<choose your app> <label>" working or suddenly "mail
user@label" and "ssh label" end up at different machines. This
is about providing constistent naming across everything.
Mark
--
Mark Andrews, ISC
1 Seymour St., Dundas Valley, NSW 2117, Australia
PHONE: +61 2 9871 4742 INTERNET: marka(_at_)isc(_dot_)org
_______________________________________________
ietf-smtp mailing list
ietf-smtp(_at_)ietf(_dot_)org
https://www.ietf.org/mailman/listinfo/ietf-smtp