Charles Lindsey wrote:
In <3E563BCB(_dot_)904AAEB(_at_)oceana(_dot_)com> Ken Murchison
<ken(_at_)oceana(_dot_)com> writes:
- usenet puts a greater restriction on the headers (although still being
RFC 2822 compliant)
There are a whole lot of things required by RFC 2822 that just do not work
in news. Some examples:
1. You can't omit that SP after the ':' in a header.
"Omitting" it is *not* _required_ by RFC 2822, nor is that any
different from RFC 1036s restriction re. 822, as noted by Ken.
2. The RFC 2822 syntax of the References-header breaks many threading user
agents.
Since it is unchanged from RFC 822, that's pretty hard to pin
on 2822. If the 1036 restrictions need to remain in effect
until software authors have had a chance to make revisions,
fine.
3. Comments in unexpected places will break some existing software.
The basis for 1036, viz. 822, allows comments essentially anywhere
in any structured field body. If 1036 didn't specify an additional
restriction, no such places are "unexpected", and such "existing
software" is broken.
4. There are some features of the Message-ID which, if strictly followed,
would totally break the news relaying mechanisms or which, if we were
forced to accomodate them, would impose unacceptable performance
penalties. Of course, these quirky cases never arise in real life, but
theoretically they could.
Message IDs are being discussed in detail.