Great Circle Associates List-Managers
(August 1995)
 

Indexed By Date: [Previous] [Next] Indexed By Thread: [Previous] [Next]

Subject: Re: Appending "Reply-To" in Outgoing Mails
From: Chip Rosenthal <chip @ unicom . com>
Date: Fri, 18 Aug 1995 22:35:08 -0500 (CDT)
To: scs @ lokkur . dexter . mi . us (Steve Simmons)
Cc: list-managers @ greatcircle . com (The List-Managers Mailing List)
In-reply-to: <1995Aug17 . 202552 . 7274 @ lokkur . dexter . mi . us> from "Steve Simmons" at Aug 17, 95 08:25:52 pm

Thanks for the comments Steve.  I don't want to create a long boring
point-by-point counter-rebuttal.  There are, however, a few comments
I'd like to make.

Steve Simmons writes:
> Summary: same as what I said earlier `*shrug*', but at greater length.

My concern is that if it's just a "shrug" then it shouldn't be done.
You seem to agree with me that the "Principle of Minimal Munging" is
a useful guide.  As I point out, this is not a hard-and-fast rule.
Rather, it sets a default:  unless there is good reason to, then don't.
If the balance between benefits and drawbacks is no more than a shrug,
then I think the principle kicks in and says, ``Don't.''

> gt> {#} Replies are directed back to gt-pfrc @
 angus .
 mystery .
 com
> gt> {#} To reply to the author, write to nickerson @
 BIX .
 com

I wrote that note to get people to think about this stuff.  As I said,
RFC-822 is deceptively tricky.  I showed, for instance, how munging
Reply-To can lose important information, making the sender unreachable.
I fear that most people who munge do not think about these things.

I find two interesting things about Steve's example.

First, it's clear that he *has* thought about the problems.  The added
tag-line prevents the loss of potentially important information.  It
helps address some of the other problems, such as reducing the "surprise"
factor.

On the other hand, the tag line confirms that the problems that I
raised are real.  Reply-To munging, for instance, breaks the reply
function.  If it didn't, the tag line wouldn't be necessary.

> chip>   o  It violates the principle of minimal munging.
> 
> True.

Oh sure...give me the one that I say was "invented to be broken". :-)

> chip>   o  It provides no benefit to the user of a reasonable mailer.
> 
> This presumes of course, that the readers of the mailing list all have
> access to a reasonable mailer.

Correct.  In fact, I *demand* that my list members have access to a
minimally reasonable mail system.  I will, for instance, drop a member
simply because their site bounces to an address in the header rather
than following the envelope.  Ultimately, we do the community a
disservice by coddling the mail systems that are broken in egregious
ways.

For minor forms of brokenness (e.g. no reply-to-group function), I
think they get to live with it.  Given a choice of forcing 98% of the
people to type addresses for reply-to-sender or forcing 2% of the
people to type addresses for reply-to-group, I'll pick the latter.

> chip>   o  It penalizes the person with a reasonable mailer in order to
> chip>      coddle those running brain-dead software.
> 
> I would argue that penalization is small to zero

But you wouldn't convince me. :-)  Breaking my "r" key is not a small
penalty.  If it was, then the {#} tag lines, clever as they may be,
would be unnecessary.

> chip>   o  Your subscribers don't want you to do it.
> 
> Disproof by counter-example, above.

Unfortunately, I suspect the people posting to comp.mail.elm asking for
a "reply-to-From-and-ignore-Reply-To" feature think Reply-To munging
is really groovy because they can reply with a single keystroke.  If
somebody would take the time to teach them what the "g" key does, I
think they'd change their mind.

> My summary -- if it's what the members want and is done with care
> and caution, it works fine.  I agree with Chip that in *most* cases,
> automatic `Reply-To: list' is a bad idea, but there are definately
> times when it is the right thing to do.

I believe your implementation of Reply-To munging is a lot less harmful
than the "strip out the old one, slap in a new one" that is performed
typically.  It's clear you've thought about the problems, and that's my
main beef:  people who munge without considering what it does.

Nonetheless, I'm kind of struck that you've gone through all this
effort just to help one poor woman stuck behind a brain dead X.400
gateway.  She must be one special lady. :-)  Personally, I'd be inclined
to tell her she was stuck typing in the address for her responses,
and teach everybody else how the reply-to-group function works in
their mailer.

> Chip, if you'd like to drop this onto your page as a comment I'd be
> honored.

Thanks.  I am saving off whatever discussion results from this issue.
I would like to make some of that discussion available to people as
a companion to the note.  (I will, of course, contact the authors
directly for permission before doing so.)

-- 
Chip Rosenthal              Old men sing about their dreams.  Women laugh and
Unicom Systems Development  children scream.  And the band keeps playin' on.
For a good time:  http://www.unicom.com/john-hiatt/
PGP key:  http://www.unicom.com/personal/chip.html


Follow-Ups:
References:
Indexed By Date Previous: Where to get help for LISTPROC
From: Dick Moores <rdm @ netcom . com>
Next: Re: Appending "Reply-To" in Outgoing Mails
From: meo @ schoneal . com (Miles O'Neal)
Indexed By Thread Previous: Re: Clobbering "Reply-To" in Outgoing Mail
From: Chip Rosenthal <chip @ unicom . com>
Next: Re: Appending "Reply-To" in Outgoing Mails
From: meo @ schoneal . com (Miles O'Neal)

Google
 
Search Internet Search www.greatcircle.com