Charles Lindsey wrote:
In <41597AD8(_dot_)5080203(_at_)erols(_dot_)com> Bruce Lilly
<blilly(_at_)erols(_dot_)com> writes:
Charles Lindsey wrote:
No, I don't think that is right. An id-right, as defined by RFC 2822[...]
You have missed the critically important fact that in all cases
discussed, the relevant RFCs refer to RFC 822 (a full Standard),
not RFC 2822 (a Proposed Standard). And RFC 822 defines the
identifiers in question as containing a local-part and a domain.
It is the clear intention that RFC 2822 is to supersede RFC 822, and that
all standards that referred to RFC 822 are to be interpreted, so far as is
possible, in the light of RFC 2822.
No, on all counts. RFC 2822 itself does not supersede RFC 822; at the
earliest, the successor of its successor might be a full Standard which
would supersede RFC 822. Standards that specifically refer to RFC 822
must be interpreted in terms of RFC 822; EBNF differs from ABNF, etc.
RFC 2822 id-right is a problem because it changes the semantics.
Good! It is an improvement.
Replacing clear, long-standing, well-understood semantics with
something which is not clearly specified, is a marked change, and
has undefined semantics is not an improvement.
Have you heard of any cases where this change has caused a problem?
I have mentioned specific issues where RFC 2822 causes problems.