From: John Stanley (stanley@peak.org)
Date: Tue Jun 18 2002 - 20:41:29 CDT
Charles Lindsey (chl@clw.cs.man.ac.uk):
>On the contrary, changing the Subject header (whether for "site policy" or
>for other reasons) is forbidden (SHOULD NOT) by Step 10.
Oh, now I see why you don't understand the problem. You don't understand
that "SHOULD NOT" isn't a prohibition, it is just a SHOULD NOT. In RFC
2119 terms, a prohibition is "MUST NOT". "SHOULD NOT" means there may
exist valid reasons to do something", but you should "carefully weigh" the
decision to do it. That's not a prohibition, that's a caution. I'm
surprised that someone who is tasked with editing a draft standard doesn't
know RFC 2119 better than that.
The prohibition on changing a Subject header based on site policy was
actually in the part of step 5 that you unilaterally replaced. Prior to
your decision to change the consensus, an injector that saw a Subject
header that violated site policy WAS prohibited from processing the
article further. You know, that part about "unaccepable" and "MUST NOT"
which you changed to "rejected". You now allow an injector to say "I won't
stop processing it, I'll FIX IT MYSELF and I don't even have to bother
asking the poster if the fixes are acceptable to him." Step 10 was never
reached.
So, now, you've decided that it is alright for an injector to change
Subject, Keywords, Summary, From, Sender, References, or any other header
that the admin has decided should be a certain way to meet "site policy".
I don't think that was the consensus of this group, but damn if you didn't
just decide that it was and change the draft without a single request to
do so and not a word of discussion about it.
>But take another example. Is it allowed for an injecting agent to insert
>an Organization header (assuming none was present beforehand)? This is
>routinely done in lots of places by way of "site policy", and noone except
>yourself has ever thought that was wrong.
Gee, I've not yet had a chance to tell you if I think it is wrong or not
and you're already saying that I think it is wrong. You just don't stop
trying to put words in my mouth, do you?
Your last sentence is provably wrong, by the way. The consensus of this
group is:
NOTE: Posting and injecting agents are discouraged from
providing a default value for this header unless it is
acceptable to all posters using those agents.
While the word used isn't "wrong", it is "discouraged", and gosh, maybe
there's a reason for that? And then we go on with this message to the
poster:
Unless this header
contains useful information (including some indication of the
posters physical location) posters are discouraged from
including it.
Now, tell me, Charles, if the injector cannot even know who the poster is
with any certainty, how the hell is it supposed to know "useful
information" about his physical location? So, yes, it is wrong, even
though it is current practice for injectors to guess about this. Since you
never seemed to notice before, maybe you will this time when I tell you
that "it's current practice" is not sufficient cause to allow stupid
things to continue. And I'll point out that the purpose of a standard is
to STANDARDIZE things, not just describe a willy-nilly ad-hoc way of
guessing about what things mean.
I'd wager that most people don't care if an injector inserts goofy or
wrong info in an Organization header, not that they think it is ok as a
standard practice. Yes, there is a difference, and no, they wouldn't
bother complaining about it. And no, bogus Organization headers do not
contradict the poster-specified From header, nor are they generally usable
as addresses for spam, so comparing them to Sender headers is specious. It
IS our business to care about such things, since we (and not just you) are
supposed to be writing this standard.
And now I'll point out yet another flaw in your argument. If we cannot say
anything in this standard that goes against current practice, then what do
we do with step 10, which says an injecting agent MUST NOT (a real
prohibition) alter the body of an article in any way? Have you missed all
those injectors that append ads to the bottoms of the articles they
inject? Altering bodies is current practice, Charles. How dare you
prohibit it!
>In fact, it appears that the
>original wording of Step 5 inadvertently prohibited this.
If the wording prohibited it, and not a single person thought it was wrong
to prohibit it, then it wasn't inadvertent.
>The new wording now allows it.
And yet you previously said that this was not a change. Is this another
one of your "unacceptable/acceptable" kinds of definitions? This is a
significant change.
> Well it is now Tuesday, and NOBODY else has asked for a straw poll.
You put a BEFORE MONDAY deadline on it, Charles.
>Therefore there will be none, and this whole issue is therefore now
>CLOSED. If you do not like this, then you will need to take it up in email
>with our Chairman.
I have already objected to your unilateral change in writing to our
chairman and have not had the courtesy of a response. He's participated in
this group since I sent that email, so I see no excuse for it.
Dave, you've got an editor that is writing the draft instead of editing,
changing the consensus at his own whim, and doesn't even know the meanings
of the RFC-standardized imperatives. When do you think you might get
around to actually paying attention to the problem?