Re: Question: UseSTD3ASCIIRules on RHS of IMA

Warning: This has gotten off topic.  This is a rathole, whose only
possible purpose is to illustrate that this topic is a rathole and
should be avoided in IMAA like it was avoided in IDNA.  :)

John C Klensin <[email protected]> wrote:

> The rule was changed in RFC 1123, which is what permitted 3COM to make
> that registration.

RFC-1123 relaxed the syntax of "host names".  RFC-821 refers to its
<domain> field as a "host name" because when RFC-821 was written, it was
always the name of a host.  But almost three years before RFC-1123 was
published, RFC-974 had generalized the domain part of a mail address
to be either the name of a host or the name of an MX record (which is
not a host, and RFC-974 always calls it a "domain name", never a "host
name").  So when RFC-1123 relaxed the syntax of "host names", did the
SMTP <domain> field even qualify as a host name anymore?  If not, what
tells us that the relaxation applied to it?  Nothing that I can see.

This may have been an innocent oversight, but I can't find any
clarification.

RFC-2821 tries to have it both ways.  On the one hand, it alters the
<domain> syntax to match the RFC-1123 host name syntax.  On the other
hand, RFC-2821 states very clearly that names of MX records are *not*
host names, which suggests that if the syntax of "host names" were
relaxed again, that relaxation would not automatically apply to the
<domain> syntax of RFC-2821.

AMC