Re: Lyris

[Vicki Richman <vicric @ panix . com>]

Chip Rosenthal has written:

> To: munging is much worse than Roger implies.  It's a
> complete abomination.

I may be wrong, but I got the impression that the feature is
aimed at one-way exploder lists, for announcements (or
advertisements) only. That is, the list-owner is also the
only message sender.

Typical exploder-list code writes the 'To: ' header
something like this:

	To: (recipient list suppressed)

Lyris, seeking cash customers, aims at greater
sophistication. Inserting each recipient into a separate
copy of the message leads the recipient to believe it is a
private message. In hard copy, it's called "mail merge"; the
most familiar example is the "You've won $1,000,000" junk
mail.

I'd guess that typical Lyris customers would be email
advertisers -- dare I say spammers? -- seeking a return on
their investment.

That feature also slows the mailing considerably, despite
the poster's claims to speed. In ordinary list messages, one
copy is sent out to multiple recipients (although long lists
might be broken into blocks of 50 or 100, to reduce
errors). But if the 'To: ' header is rewritten for each
recipient, a separate copy is sent to each recipient.

There is no way multiple copies can match a single copy to
multiple recipients in speed.

A year or so ago, I wrote a tcsh shell script that did
precisely what the Lyris poster claims. I was sending
occasional -- fewer than one a month -- announcements to
about 1000 dues-paying members of my organization. (A
'while' loop invoked /usr/lib/sendmail for each rewritten
copy.) Even though I labeled the mailing as public in a
user-header, a large number of recipients wrote back
thanking me for what they believed was a personal message.

Mailing to the list of 1000 took about 90 minutes, with my
Unix shell script. Compiled binary code could do the job
many times faster, but I doubt that hundreds per second
could be achieved. Maybe ten per second. Maybe.

I quickly realized that my script could be used for great
evil as well as good, and I stopped using it. If I wanted to
identify each recipient separately, I'd now use a
user-header, say 'X-Recipient: '. To avoid deception, the
'To: ' header must identify the mailing as to a public list.

> The unstated implication is that Reply-To is munged as well.
> I'm jumping to that conclusion because I don't see how message
> r)eply works since the To: header has been corrupted.
> 
> I'll elide my usual Reply-To munging rant.  Check out
> <http://www.unicom.com/FAQ/reply-to-harmful.html> if you
> haven't heard it yet.  I will, however, quote the one
> passage that says:

Chip, I read your 'Reply-To: ' FAQ, and I disagree
completely. My lists and a few of my favorites direct
replies to all. I have taken to inserting a footer
instructing subscribers how to rewrite the headers for a
private reply.

I used the Elm 'g' option for this reply, and Chip, you're
getting two copies. That would annoy me. I did it just to
make my point. Ordinarily I would make a note of the list
address, hit 'r', then hopefully remember, before hitting
's', to remove the individual from the 'To: ' line and to
insert the list address in its place.

I believe Listserv allows the sender's 'Reply-To: ' header,
if it exists, to take precedence. If it doesn't exist, as it
almost never does, then Listserv inserts the list address
into that field. I like that. It gives the sender the option
on directing replies.

----------------------------------------------------------------------------
Vicki Richman             vicric@panix.com            National Writers Union
Harlem, New York              PGP 2.6                UAW Local 1981, AFL/CIO
"You are about to witness an experiment: the effect of electricity on Wood."
     -Frederick Wood, the last person electrocuted by the State of New York.
----------------------------------------------------------------------------