[TIP] Reply-to settings (was: Re: loading test data into frameworks)

C. Titus Brown ctb at msu.edu
Fri Jan 14 13:43:51 PST 2011

> [ please set ?the reply-to to testing-in-python@ !]

As you may not know, this is actually quite a contentious subject with a long

The topic first came up with the rise of ARPANET, in the early '70s.  As e-mail
grew, people naturally wanted mailing lists so that they could send messages
more quickly and easily to large, largely irrelevant groups of people.

There quickly e-merged a divide between those who wanted replies to
automatically go to the list, and those who wanted such a choice to be made
intentionally with each message.  Adherents of the former setting, known as the
Apathetic Autorepliers, or "appies", and the latter group, Intelligent
Intentionists, or "innies", quickly came to (metaphorical) blows.  The former
group spoke glowingly of the implicit communities formed by such
mailing lists; the latter group decried the inevitable spam and apologetic
follow-ons stemming from the configuration.

The late '70s were characterized by a number of sharp incursions and repostes
by both groups, culminating in the November, 1979 meeting mediated by Eric
Allman, author of delivermail and sendmail.  Allman -- who, as the author
of mail transport software, had little invested in the topic -- moderated
a rancorous debate between, among others, Dave Taylor and Steve Dorner,
each of whom insisted that they were correct in their stance and decried
their opponents views.  This debate was famously summarized by Eric Allman
with the text


which, when run through an early version of the m4 preprocessor, expanded to a
243 page treatise outlining both positions.  (Perlers may recognize this text
as Larry Wall's first test case for Perl v1 syntax.)

By the '80s, the lines were clearly drawn and entrenched forces began to build
a dividing line in the form of the Internet e-mail software ecosystem.  The
rise of CompuServe and AOL confused matters here, however, because appies and
innies were unable to get CIS or AOL to commit to any standard; and, far from a
single, cohesive approach, the innies and appies started to fragment as CIS and
AOL dominated Internet usage.  It's interesting to note here that early
versions of smail and qmail were born during this period, but their release was
delayed due to the intense partisan fighting over the SMTP standard.

The Hazel Accords of January, 1992, hosted by Bush Sr. at Camp David, led
to the first clear statement: either configuration setting was considered
formally acceptable, and institutions were encouraged to declare themselves
"Inny" or "Appy".  (This usage has since been deprecated, due to the
continual threats of lawsuits from Apple.  Go figure.)

John Viega, Ken Manheimer, and Barry Warsaw clearly knew of these accords when,
in the late '90s and early 2000s, they developed mailman.  Hence there is
a shiny little checkbox entitled "Where are replies to list messages directed?"
Unfortunately for concordance, Barry Warsaw chose to take an ideological
approach, that of an inny -- immediately following the text, there is another
sentence, "Poster is strongly recommended for most mailing lists." Even more
unfortunately, Warsaw's official position as FLUFL means that this decision has
percolated throughout the Python community, with only token resistance offered
by the BDFL, who appears to be entirely under Warsaw's spell.  Thus, both
Mailman itself and all of the core Python mailing lists follow an "Inny" policy
by default.  While I was not actually active at this time myself, Terry Peppers
may serve as a useful resource to this period of ancient Internet history.

In recent years, there has been a general die down of the fight between
"inny" and "appy" partisans.  This is probably due to a combination of the vast
increase in the number of Internet users, most of whom don't give a damn; this
is best exemplified by the quote, "I don't give a rat's ass about... what were
we talking about, again?", by Steve Holden, referring to the imminent release
of Mailman 2.1.15 in January 2015 (newly ported to Python 3!)

So, I'm sure you can appreciate the rich history of this discussion, and
I hope that agree -- whatever your side -- that there is no really *right*
configuration option.  I have chosen to go with the FLUFL's position here,
and, while I acknowledge that other people have the right to their own
opinion, they are actually just wrong.

In closing, I refer you to the briefer discussion here,


...  Therefore, if the mailing list management software offers a way to
automatically cause discussions to stay on the list, you would think turning
that feature on would be the obvious choice.

Well, not quite. There is such a feature, but it has some pretty severe
disadvantages. The question of whether or not to use it is one of the hottest
debates in mailing list management - admittedly, not a controversy that's
likely to make the evening news in your city, but it can flare up from time to
time in free software projects. ...

and plead for your forbearance on, if not your understanding of, my
configuration choice.

[ speaking for the 43-strong testing-in-python administrative team ]
C. Titus Brown, ctb at msu.edu

More information about the testing-in-python mailing list