From: Charles Lindsey (chl@clw.cs.man.ac.uk)
Date: Fri Sep 29 2000 - 05:05:07 CDT
In <ylwvfwnqy7.fsf@windlord.stanford.edu> Russ Allbery <rra@stanford.edu> writes:
>I think more to the point this is outside of the scope of an IETF RFC. We
>can indicate the meaning of the headers, we can state what automatic
>program behavior should be, but when it comes down to people explicitly
>and intentionally doing something contrary to the stated meaning of the
>headers, there really isn't all that much that we can do. It's only a
>technical interoperability standard, not a law.
Indeed. The most we can do is to write it in such a way as to imply that
such bahaviour is a Bad Thing (by writing Ought, or otherwise).
>Accordingly, I'd prefer to avoid any mention at all of what the software
>should do if the user explicitly overrode it, since at this point we're
>talking about user interface issues that seem rather outside the scope of
>the standard. Instead, I think we should state that if Mail-Copies-To is
>set to "never", this indicates that the poster never wants an e-mailed
>copy of followups and that a followup agent MUST NOT send such a copy. Is
>there any real need to say anything more than that, or even mention user
>overrides at all?
No, if you say that, some over-zealous jobsworth will interpet it as
requiring all posting agents to detect that a poster is emailing a copy of
a document which just happens to be identical to an article he has just
posted AND that the person he has emailed it to is some complex alias for
the poster of the original ... :-( .
I think SHOULD NOT covers it adequately (it means you certainly don't do
it automatically, but you don't have to bloat your implementation to cope
with the bizarre ways users might try to get around it).
>If people want to force a copy for some bizarre reason (and I can think of
>times where it may be necessary, although all my examples are within
>cooperating subnets), well, then they're not using a followup agent any
>more.
The exact difference between a followup agent and a posting agent that
just happens to be following up is one we don't want to get drawn into.
Using a wording such as "explicitly overrides" means that the distinction
does not really matter.
>Similarly, for the case of no header, why not just say that if the
>Mail-Copies-To header is not present at all, the article should by default
>be treated as if it had "Mail-Copies-To: never"? This neatly leaves open
>the possibility of changing the defaults for some particular circumstances
>(such as mailing lists gatewayed to newsgroups, like many of the gnu.*
>lists, where you have to send a copy to the original poster under many
>situations or they'll never see it -- they're often not aware that the
>address to which they're mailing reaches a newsgroup at all), while still
>making the general protocol statement we want to make about the common
>case.
I think the present wording is just about right:
If this header is absent, a followup agent MUST NOT automatically email a
copy of the followup article, but the user MAY do so manually, at his
discretion.
I have alterd this to read:
If this header is absent, a followup agent MUST NOT automatically email a
copy of the followup article (but this does not prevent the user doing so
manually, of course).
-- Charles H. Lindsey ---------At Home, doing my own thing------------------------ Email: chl@clw.cs.man.ac.uk Web: http://www.cs.man.ac.uk/~chl Voice/Fax: +44 161 437 4506 Snail: 5 Clerewood Ave, CHEADLE, SK8 3JU, U.K. PGP: 2C15F1A9 Fingerprint: 73 6D C2 51 93 A0 01 E7 65 E8 64 7E 14 A4 AB A5