[Top] [All Lists]

Re: Dotless domains

2009-02-07 14:47:19

I'm getting very itchy feet here.

On 25 Jan 2009 14:31:59 -0000, John Levine <johnl(_at_)taugh(_dot_)com> wrote:

IMHO, there should be an RFC explaining that use of a TLD as a
terminal domain name was never intended to work, and should not be
expected to be supported by the Internet architecture or software.

Sounds good to me, although there is the pesky detail that 2821
ambiguously and 5321 unambigously says that they do.

I do have a problem with the "DNS is entirely transparent" attitude
maintained by some people, just as I have issue with some other
peoples' views that SMTP is eight-bit clean, just because they want it
that way.  Here again, and I felt sure we'd dealt with this by now,
the question of interoperability is the deciding factor.

For me, RFC 5321 in the matter of dotless domains is as it should be
and (textwise, at least) as in the predecessor makes an appropriate
allowance.  It has been, and always was, how my MTAs have been
configured, and for me to configure an MTA otherwise would be to
configure a flawed MTA on the Internet.

So, curse me blind (I already am, so you can do that, you see) for
actually supporting the notion that, regardless of known issues, we
should favour consistency between two different systems with different
historical (hysterical?) problems.  I'd never do this for the
"Underscores-in-hostnames" thing (I don't have them), and we didn't do
it for the IPv6 implicit MX, and I'm pleased we don't continue to
advise obviously bad ideas with respect to SMTP.  But, it's not
outside practical implementation to keep the current semantics with
the specifications as written.  If you do make a document reminding
everybody of the historical context for SMTP with respect to a flat
namespace, please make it advisory only.  I have no problem with
technical implementations continuing to do what RFC 2821/5321 say they
can, and (if we have to put up with this newfangled TLD-a-penny
scheme) I'd sooner we be prepared than standoffish.  I only wish the
people who point to new specifications as excuses for breaking old
ones could please kindly stop it, thank you very much.  Of course, I
don't think RFC 5321 is an example, and I don't think the author had
thought of the historical FQDN/hostname/domain terminology and was
simply spelling out what would have been obvious to implementers of


Sabahattin Gucukoglu <mail<at>sabahattin<dash>gucukoglu<dot>com>
Address harvesters, snag this: feedme(_at_)yamta(_dot_)org
Phone: +44 20 88008915
Mobile: +44 7986 053399

<Prev in Thread] Current Thread [Next in Thread>
  • Re: Dotless domains, Sabahattin Gucukoglu <=