From: John Stanley (stanley@peak.org)
Date: Tue Jan 20 2004 - 21:09:59 CST
Bruce Lilly <blilly@erols.com> writes:
>This is far too much verbiage. It suffices to say:
>The Subject field is unstructured. No assumption should be made
>regarding its content.
Charles Lindsey (chl@clerew.man.ac.uk):
>We have discussed this extensively before. I think you are alone in
>demanding this minimalist position.
Subject-header = "Subject" ":" SP Subject-content
Subject-content = unstructured
2* unstructured = 1*( [FWS] ( utext / encoded-word ) ) [FWS]
2 utext = NO-WS-CTL / ; Non white space controls
%d33-126 ; The rest of US-ASCII
encoded-word = "=?" charset ["*" language] "?" encoding
"?" encoded-text "?="
As I recall, [x] is optional, (a/b) means "a or b", and 1* means "one or
more".
Either the header is unstructured or it is not. If it is unstructured,
then it is exactly correct to say that no assumption should be made about
its content.
>Others were adamant that some mention
>should be made, at least to the extent that "if you do this, then you
>SHOULD do it properly so that is can be recognized reliably".
There is no means to recognize structure in an unstructured header --
that's what "unstructured" means. There is also no need to create
structure where none exists: there is already a MANDATORY header provided
which not only tells you, by its very presence, that this article is a
reply, but points back specifically to what article it is a reply to. This
header is MANDATORY and has been for ten years.
There is nothing reliable in the "Re: " hack -- we've already gone over
the many ways that it can be present without the article being a reply.
I'm sure we all remember them. "Subject: Re: considered a hack" is just
one of the many.
So, both "recognize" and "reliable" are meaningless in this context. Is
this draft the correct place to formalize meaningless and obsolete
kludges, or should such things be elided?
I don't think Bruce is alone in this, and if you remember the previous
discussions, you'd not think it, either.