ietf
[Top] [All Lists]

Re: precedence field and mailing lists

2000-07-13 14:50:03
From: Doug Royer <Doug(_dot_)Royer(_at_)SOFTWARE(_dot_)COM>

...
Could you please point me at the standards (or otherwise) documents
that describe the semantics of 'Precedence:'?  It sounds interesting
but I just can't seem to find out what you mean or expect it to do.

`man vacation` on any modern BSD-like UNIX box should say something like:
   No message will be sent unless login (or an alias supplied using the -a
   option) is part of either the ``To:'' or ``Cc:'' headers of the mail.  No
   messages from ``???-REQUEST'', ``Postmaster'', ``UUCP'', ``MAILER'', or
   ``MAILER-DAEMON'' will be replied to (where these strings are case insen-
   sitive) nor is a notification sent if a ``Precedence: bulk'' or
   ``Precedence: junk'' line is included in the mail headers.  

Near the end you'll likely find something like:
   The vacation command appeared in 4.3BSD.
which the knowledgable translate into "official, open source publication
in 1985 or 1986, depending on access to the Beta tape."
Then there are the at least as ancient words in sendmail documenation.


...
In other words, the IETF itself is not above some standads bending..
what standard are you referring to?
How about hoping that X-Loop will be preserved by those peered mailing
lists and other gateways and so prevent loops?

FYI: X-<anything> is by definition - not standard header.

While your words agree with my point, you say them as if you disagree.


...
A standard is what http://www.m-w.com/ says, "something established by
authority, custom, or general consent as a model or example."

...
Something can be 'standard' and never used by anyone. Something can be
used by everyone and never a 'standard'. 

Perhaps according to the Church of De Jure Standards.
Among those who disagree are dictionaries, heathens, apostates,
and others whose livelihoods or self-esteem don't require the
validation of what other standards bodies officially and revealingly
call "The Standards Process."

...
I'd rather that AOL took outside comments on their standards, but they're
still standards.

Please point me at those standards. Or did you mean undocumented
protocols? Or did you mean documented and not public protocols?

I mean AOL's protocols, applications, and other computer stuff which meet
the definition of "standard" in http://www.m-w.com/ by being "established
by authority, custom, or general consent as a model or example." 
The consent and custom of 23,000,000 people and the authority of AOL meet
all three of those dictionary requirements.  They don't and should not
meet modern IETF requirements for standards track RFCs, but I suspect that
if (perhaps a big 'if') they met technical muster, they would have been
accepted 18 or 20 years ago by the ad hoc group that preceded the IETF.

Which gets me back to the Precedence: header.  That some people dislike
it, many more are ignorant of it, and it has more than one use does not
imply that the IESG should not be using it as the best and only available
fix to a real problem.  That the Precedence: header meets the dictionary
definition of "standard" is gravy, except for the awkwardness of the IETF
admitting that the IETF is not the font of all wisdom.  Long ago that
self-knowledge was a given, and at still is among "legacy programmers."


...
I think the standard definition of standard includes, approved,
reviewed, and publicly available.

That implies you think that TCP, IP, STMP, and the rest of the IETF's
claim to fame are not standards because they were not externally (since
I'm sure AOL reviews AOL stuff) reviewed more than AOL's, were in practice
no more publically available (who knew of them?), and several years after
being specified, were not merely unapproved but explicitly forbidden by
many national goverments in the many GOSIP's.

Believe it or not, SMTP was not created this year with the publication of
the new SMTP RFC, nor even 1989 with the minor elaborations in RFC 1123.
TCP is not significantly changed from it's ancient, unapproved, unreviewd
(by outsiders), not publicly available (because of obscurity) form.  Thus
the only large scale transport protocols available that meet your
particular definition of "standard" are TP0 through TP4.

Why do I suspect some of the most vigorous supporters of the sanctity of
standards committee authority don't know what I mean by GOSIP and TP4?
In other words and as I said before, many of the revolutionaries moved
into the deposed overlords' offices and accepted fealty from the toadies,
go-ers, empire builders, and courtiers of the old regime, and things went
back to normal.


Vernon Schryver    vjs(_at_)rhyolite(_dot_)com



<Prev in Thread] Current Thread [Next in Thread>