From: Bruce Lilly (blilly@erols.com)
Date: Tue Jun 29 2004 - 22:50:32 CDT
Frank Ellermann wrote:
> Eivind Tagseth wrote:
>
>
>>The ABNF would look something like this:
>
> [...]
>
>>Which I find quite messy.
>
>
> That's true, but if you look for some of the obs- rules in
> RfC 2822... it's not better, or even worse.
That is likely to change in the 2822 successor -- it has been discussed
on the ietf-822 mailing list (q.v. if you are interested).
> We could use a mandatory "Re: " rule only for obs-subject,
That's been tried; it doesn't work and has been abandoned.
> The first link was a statement about new-subject, and all
> "Re: " and body conventions derived from RfC 2822 belong
> to USEAGE (or as Bruce said, only "over-the-wire" issues
> are relevant for USEFOR. The side-effects of obs-subject
> cmsg _are_ "over-the-wire").
They seem to be "local policy" issues to me, as I've stated.
We *do* need to address the 180-degree reversal relative to 1036,
but that's another matter.
> The second link says "no new ABNF for 2822 subject", and
> "document existing practice". That's IMHO a contradiction,
> obs-subject _is_ existing practice. If Alexey wants to
> solve this problem by decree "RfC 1036 servers are broken",
> now that wold be... interesting.
In reality, you're talking about *a* subset* of servers; I
wonder what proportion actually honor cmsg.