This is just too rich. It's now come up twice, within a space of
a couple of hours! You've tried to tell us, twice, about this one
unlikely string or pattern that no one would (or should) ever
type, but the list software has replaced it with the empty
string, so we're still in the dark!
Josh, I'm afraid you're going to have to concede defeat on this
one. There are some extremely cynical old curmudgeons on this
list, and after seeing the number of e-mail addresses that got
substituted before the feature got turned off, and the mass
confusion engendered by the substituted e-mail addresses, and now
the cryptic reference to this new secret string that we can't use
and that's so secret you can't even tell us what it is, there's
just no possible statement you can make on the subject of message
body rewriting that people here are going to find acceptable
other than "message body rewriting has been turned off."
I'm sorry that people (myself included, in my first paragraph)
have been picking on you for the faults of a piece of software
you didn't write, and I thank you for remaining civil and not
taking it personally. I hope you can laugh (as I am) about the
whole issue (and especially the mystical substitutions);
it's all really quite funny.
> The outstanding issues and their status that I'm aware of:
> 1) Invalid Message-IDs
> Currently deciding on the exact fix. Will still be a new message-id, but it
> will at least be a valid format. Should be fixed in upcoming version.
I'm still not convinced that each copy sent out needs its own
message-id. *IF* each message were different, I'd understand,
and I'd immediately propose the same explanations you've already
given -- distinct messages clearly need distinct message ID's.
But the reasons for individual recipients' copies of list
messages to be distinct were: (a) e-mail address rewriting, which
will be turned off; (b) other rewriting, which I believe I've
established will have to be turned off; and (c) footer appending,
which I think we've agreed this list doesn't need. In other
words, once we have the new list software configured the way we
want, all copies of a particular message sent out will be
identical, after all, so there's no reason they can't all use the
same message-id.
(There's anther incentive to keep a single message-id, which
hasn't been mentioned yet. If a followup message makes reference
to a previous message -- either in the In-Reply-To: or References:
headers, or in somewhere in the body of the message -- it's of
course natural to use the message-id. But no existing mailer
software is ever going to construct an In-Reply-To: line based
on Lyris's new X-Message-Id: header, which we can of course
immediately tell is nonstandard by the leading X-. Take a look
back at the In-Reply-To: headers for the messages to this list
since the software changeover -- every one of them that mentions
a message-id is meaningless to all but one of the members on the
list.)
All in all, I have to say I'm puzzled about why the Lyris
software chose the approach it did. It's true, as Al Gilman
pointed out earlier, that "`How do I get off?' is FAQ #1 on any
list... where something akin to this footer hasn't been the list
practice". But rather than go to all of this trouble --
appending footers, rewriting e-mail addresses, generating unique
message-id's, and finally trying to justify these decisions to a
skeptical audience -- why not just scan all messages on their way
to the list for strings of the form "please remove me", and
divert them for manual processing? I believe some list software
does this successfully already, and without strong AI or
anything.
Steve Summit
scs@eskimo.com
[
Usenet Hypertext FAQ Archive |
Search Mail Archive |
Authors |
Usenet
]
[
1993 |
1994 |
1995 |
1996 |
1997
]
© Copyright The Internet FAQ Consortium, 1997
All rights reserved