From: Charles Lindsey (chl@clw.cs.man.ac.uk)
Date: Mon Apr 23 2001 - 11:44:04 CDT
In <Pine.LNX.4.10.10104131701020.15822-100000@spock.peak.org> John Stanley <stanley@peak.org> writes:
>Charles Lindsey (chl@clw.cs.man.ac.uk):
>>Because new headers are going to appear in the future,
>I am not talking about the future. I am talking about right now. MIME
>headers are already defined, yet THIS DRAFT says that software SHOULD NOT
>attempt to interpret them.
NO IT DOES NOT!
Forgive my shouting, but some people seem to be deaf. Read 2.4:
... Certain syntax from the MIME specifications [RFC
2045] et seq is also considered a part of this standard (see 6.21).
and in 4.2.1, right at the start of the paragraph you are complaining
about:
Header-names SHOULD be either those for which a USENET-header-content
is defined in this standard, or those defined in [MESSFOR], or those
defined in any extension to either of these standards including, in
particular, the Mime standards [RFC 2045] et seq., ...
and in 6.21:
The following headers, as defined within [RFC 2045] and its
extensions, may be used within articles conforming to this standard.
So those headers are certainly "described in this standard" for the
purposes of 4.2.1 (indeed, they are discussed extensively in 6.21).
Now I might be persuaded that the words in question need some change, but
NOT for the spurious reasons that you are claiming.
>>>If I type
>>>my name at the end of an article, is this a "signature" as defined by this
>>>standard and am I violating it by not jumping through the hoop?
>>Yes.
Hmmm. I did not make myself very clear. 4.3.2 defines the term "personal
signature" with a very specific meaning (notably it comes at the end of a
body or Mime part and it has '-- ' in front of it). Indeed, the '-- ' MUST
be present in order to satisfy the definition.
Now if you write your name, your address, and other stuff at the end of
the article without a '-- ', then it may be a "signature" (in common
parlance) but it is not a "personal signature" (as defined in our draft).
Nothing wrong with defining such a thing in our draft, saying how it is to
be constructed, and what various agents Ought to do with it. However, it
might be argued that such a specific concept requires a more distinct name
than "personal signature". Would "Usenet signature" be a better term?
>To claim by use of a MUST that I am creating some interoperability problem
>by typing my name at the end of an article is patently absurd. ...
There are several places where we use MUST to specify how to contruct
objects without an explicit syntax (as here) or to impose additional
syntactic requirements or to resolve sysntactic ambiguity (as with "Re:
..."). Interoperability arises to the extent that software which attempts
to recognise (or parse) such objects and ignores those MUST requirements
will then go on to do something wrong. There is a similar use of MUST to
resolve syntactic ambiguity in the draft syntax for wildmats in the NNTP
draft, and nobody there is claiming that it is an improper usage.
>>If someone can tell me a good nroff hack to make it do that ...
>Do as we say and not as we do... Did you actually NOT include the
>mandatory space in the copy you sent IETF?
I think IETF drafts and RFCs are meant for human reading. If some
non-humanly detectable feature (such as trailing whitespace) manages to
get through to the stored form of the final document, then that is a
bonus. But there is no guarantee that it would then make it to other
copies of that document, so one must not rely on it (and we do not rely on
it).
If there is some convenient hack that would make it appear, then I would
try to do so, but I do not think it worth exceptional effort to get it
right.
>>The fact that some injecting agents are not as omniscient at their authors
>>would believe does not alter the fact that some of them DO attempt such
>>stupidities.
>Yep. And it is stupid for us to admit we know it happens without fixing
>it. That's as much as a stamp of approval.
>> I presume it was a
>> real problem that had been observed in practice.
>Then the text should be a restriction on injecting agents to prevent it
>from being conforming behaviour. When it is a MUST NOT, then it is not an
>injecting agent, it is rogue software, when something tries doing it
>wrong.
>"The requirement for the From header is any address belonging to the
>poster. Because injecting agents cannot determine which addresses meet
>this requirement, they MUST NOT attempt to enforce it by replacing user
>provided entries with their own guesses."
One could strengthen the requirements on those lines. OTOH, one might
argue that what an injector does it a matter of negotiation between its
owner and the users that use it. If you don't like what your local injector
does to your articles, then find someplace else to inject your articles.
OTOH, now that we have the Injector-Info header, there is less reason for
injecting agents to do such munges. So does anyone else what a change in
that area? To remind you, the present text is a NOTE:
Be warned also that some injecting agents that have
authentication information may choose to replace the From-
content based upon the authenticated identity.
Are such agents really still around?
Possibilities are to remove that NOTE, to leave it alone, or to outlaw
such behaviour. Opinions?
-- Charles H. Lindsey ---------At Home, doing my own thing------------------------ Tel: +44 161 436 6131 Fax: +44 161 436 6133 Web: http://www.cs.man.ac.uk/~chl Email: chl@clw.cs.man.ac.uk 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