[erlang-questions] 'reply-to' header in this mailing list

Michael Turner <>
Wed May 18 10:36:43 CEST 2011

Pierpalo asks: "According to your intuition, what should be the difference
an action called "Reply" and another called "Reply to all"?"

My intuitions about what ANY response should be to ANY request for action
depend on my personal experience and on the circumstances of the request.
And I'm hardly alone in that. In user interface design, this is called "The
Principle of Least Astonishment."


<http://en.wikipedia.org/wiki/Principle_of_least_astonishment>On a mailing
list, I'm slightly astonished ("dismayed" would be a better word) when I
reply to the list (or so I think) and discover later (IF I ever do) that the
reply went to an individual instead. Here, I would expect "Reply" and "Reply
to all" to do the same thing. That's redundant, you say? Well, so what? Life
is redundant. Then you die.

I understand the argument from risk: Yes, every so often, somebody thinks
they are replying to an individual when in fact they are replying to the
list. I've done this mysefl, and suffered embarrassment.

But risk correlates with reward, as with anything in life, and a mailing
list is *always* a balance between individual risks of embarrassment ("Will
I seem stupid to some people? Will the answer to my question come to me a
second after I hit Send? Was there an embarrassing typo in what I wrote?")
and collective reward: everybody gets to see the answers; the only people
who lose are those who find the answer--or the inbox clutter--annoying.
(Presumably, the signal averages out to be positive for most; and for those
for whom it doesn't, unsubscribing is always an option.)

So, in this case, doing the more expected, less astonishing thing is quite
aligned with the spirit of mailing lists, even if it's in violation of the
letter of IETF law. Sometimes, the most rational thing to do is not the
purely logical thing to do. In those cases, I side with rationality. But
look: there's cultural variation. Maybe most people in the Erlang world have
different expectations or thresholds of "astonishment". So the most rational
thing would probably be to put it to a vote, on the list. After all, the
issue's just going to come up again, chewing up yet more of that "bandwidth"
that Raimo's so worried about conserving. So why not just find out what most
people want and do that?

-michael turner

On Wed, May 18, 2011 at 4:38 PM, Pierpaolo Bernardi <>wrote:

> On Wed, May 18, 2011 at 07:35, Michael Turner
> <> wrote:
> > Does anyone have any statistics on this question? That is, what
> percentage
> > of mailing lists "do it wrong"? It's counterintuitive to me that  a
> "reply"
> > on this mailing list is only to the individual. "Reply" is "Reply to the
> > list" on every other mailing list I'm currently on, and on almost every
> > mailing list I can remember being on. But perhaps my lifetime mailing
> list
> > membership doesn't approach statistical significance.
> > Although it may get me accused of "blindly following the herd," let me
> > assert it anyway: arguments from "intuition" in commonly used software
> > interfaces only work when you have statistically significant user support
> > for them, not some purely formal, rule-based argument for your *personal*
> > intuition. What's "intuitive" to one person may be counterintuitive to
> many.
> According to your intuition, what should be the difference between
> an action called "Reply" and another called "Reply to all"?
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://erlang.org/pipermail/erlang-questions/attachments/20110518/4198e01a/attachment.html>

More information about the erlang-questions mailing list