ietf-smtp
[Top] [All Lists]

Re: domain name definition in RFC2821

2006-04-22 00:59:01



--On Saturday, 22 April, 2006 02:42 -0400
Valdis(_dot_)Kletnieks(_at_)vt(_dot_)edu wrote:

On Fri, 21 Apr 2006 21:44:03 PDT, Yuri Inglikov said:
There appears to be some disconnect between ABNF syntax and a
prose. I.e. i t appears that ABNF requires at least 2
sub-domain parts, while prose discu sses "one or more
dot-separated components". Which one is "more correct" an d
any scenario when a single-component domain name can be valid
/ useful in modern SMTP?

Can you give an example of a "single-component domain name"
that would *not* be flagged as a failure to canonalize to a
FQDN?

Are there any e-mail addresses that *work* (or even could
*potentially* work) of the form 'userid(_at_)com' or 'userid(_at_)net'
or 'some(_dot_)full(_dot_)name(_at_)to' or anything else like that?

While it has not been common practice, nothing in principle
prevents 
    Random-TLD.  NS server-for-Random-TLD.Random-TLD.
                 NS other-server.other-TLD.
                 MX mumble.somehost.some-TLD.
Knowing what I do about SMTP implementations and about the
problems, especially "search rules", you cite below, I'd
strongly recommend the use of 
   poor-sod(_at_)Random-TLD(_dot_)
rather than
   poor-sod(_at_)Random-TLD
but...
 
(And even if you require 2 components, it's traditionally been
ugly.  There were a *lot* of university computing centers and
comp-sci departments that were quite surprised when all the
addresses of the form 'userid(_at_)server(_dot_)cs' or 
'userid(_at_)server(_dot_)cc'
were no longer recognized as a shorthand for
'server.cs.foo.edu' or 'server.cc.bar.edu'.  And some of us
were around for when British schools leaked reverse-order
domain names, and 'user(_at_)uk(_dot_)ac(_dot_)FOO(_dot_)cs' went to another 
country
rather than to U of Foo's CS dept...)

Yep.
    john