[OSM-talk] no replay-to header at talk

Dirk-Lüder Kreie osm-list at deelkar.net
Wed Jul 11 01:15:53 BST 2007

Hash: SHA1

Ulf Lamping schrieb:
> Matthew Newton schrieb:
>> Hi,
>> On Tue, Jul 10, 2007 at 02:00:10PM +0200, Raphael Studer wrote:
>>> I've noticed that on the talk at osm there is no replay-to field in the
>>> header, if an email is sent to the list.
>>> On the german list (talk-de at osm) there is such a field.
>>> I would be happy if it could be added on the talk list also.
>> This request surfaces every now and then and, 
> Well, I work with mailing list of both using "reply-to" and not doing so 
> (I'll call it "hard-core") for many years now.
> Interestingly, while reading this request repeatingly at the hard-core 
> lists, I *never* have read even a single request to switch off 
> "reply-to" on *any list* I'm reading for years now. So the interest in 
> "reply-to" to be switched on seems to be much more tempting than to 
> switch it off once it is used.

It's easy to miss that there is a reply-to munging in place once you're
accustomed to using reply-all and seldomly answering ppl in private.

I didn't notice it for the talk-de list at all for example.

> While this of course is not "hard evidence", one should at least see 
> this as good reason to think about this topic in a different light.
>> sorry, no, it's not
>> going to change. There are several reasons, not including that many
>> people on the list don't want it changed; someone has already
>> pointed to a good document listing most of them.
> Sorry to say, but this document is mostly garbage!
> "Minimal Mungling": True, but only in principle. It seems, that a good 
> amount of mailing lists can handle this without any problems for years 
> now. So there's no rocket science to bring this feature to life. At some 
> time in the past even attaching anything to mails was vodoo, so should 
> we still avoid any mail attachments because they complicate the mails? 
> And believe me, I've read a lot of the corresponding RFC's about 
> attachments, this isn't really simple stuff.

So you're arguing there is no real technical need, wether to allow
Reply-to-munging is just a matter of opinion?

> "It adds nothing": Simply untrue. It adds a feature that many users 
> simply seem to expect - pressing reply will send a mail to where they'll 
> expect it to. And as I subscribed to a mailing list, replying to mails 
> I've received from it will get back to this list and nothing else.

What they expect is solely a matter of education. I for one would not
expect the reply to go back to the list if I hit "reply".

> "Freedom of choice": Utterly untrue. The sentence should be: "All 
> responses as default should go to where the user expect it to go to.". 
> Again, there seems that most users expecting it to go back to the ML and 
> *not* to the originator - by default.

While I do disagree with some points in that referenced document I do
like the choice. I am very used to use "reply all" when mailing to a
list, and "reply" when wanting to reach the author of the mail
privately, without having to manually copy over the e-mail-address.

Having no serverside supplied "reply-to:" gives *me* that choice without
additional effort.

> "Coddling the Brain-Dead, Penalizing the Conscientious": Yes, 95% of the 
> internet users are simply using the wrong software. If the linux/unix 
> community keeps this elitist attitude, Microsoft will keep it's over 90% 
> market share still in centuries. Sorry to say, but it's just that simple 
> Interestingly, a few lines later he admits, that he stopped using 
> "reply-to" mungling after sending a private mail to one of his public 
> lists. And that one calls other people "brain-dead" as they use the 
> wrong mail clients - well, I find this attitute interesting.

This isn't a linux vs windows thing, it's a "I care" vs "I don't care"
thing. Yes it should work without giving it much thought, the machine
should adapt to the user, not the user to the machine. But Microsoft is
not part of that machine, there are many e-mail clients for Windows that
*do* support Standards-compliant Mailing lists (Thunderbird being just
one example).
Yes I agree that you cannot expect users to change their software for
every other reason, but if it's broken, why keep it? Just because it
came pre-installed?

> "Yes, on occasion somebody mistakenly responds directly to the author of 
> a message when they wanted to reply to the group"
> Well, I do it occasionally, still after years. And when I send a request 
> to the list, I'll sometimes even get more direct replies than through 
> the list server - obviously not intended as there was no "private" 
> content. Seems I'm not the only one making this mistake "occasionally" ...

I did it occasionally, too, but that's no real harm, is it? I think it's
better for the default to be the safe one.

> The sentence I've found most interesting was: "I contribute to the Elm 
> mailer development team". Yes, even the start to use GUI's was basically 
> the wrong step. The whole document reminds me of Douglas Adams: "some 
> even regard that leaving the trees was the first wrong step" (roughly 
> remembered).

This implies the reply-to munging would be called progress, I doubt
that. For me it would be a step backwards.

>> Two things to think about:
>>   The list changing the Reply-to header messes things up for users
>>   who specifically set a reply-to in their own mail because they
>>   want replies to go to a different address. Reply-to is an
>>   Originator Field as specified by RFC2822, and as such should not be
>>   touched by intermediate systems.
> Some notes:
> - how many users actually uses the "reply-to" while sending their own 
> mails as described in RFC2822? I guess only very few - and will it be 
> essential if they cannot do so?
> - I guess a vote would result in a 0.001% (actually using "reply-to") 
> vs. 90% (sending their mail unexpectedly at least sometimes only to the 
> originator) :-)))

You forget the case where a definitively private mail lands on the list

>>   Your mail client* should have (at least) two options: "reply" and
>>   "reply to all". It may even have "reply to list". If it doesn't
>>   have these then get a new mail client. ;-
> Yes, I know, I don't have the right mail client. I should learn emacs 
> and elm and forget about any GUI interface and these other strange stuff 
> invented in the last 20 years.

There are fine GUI clients who allow you to use all features of a
mailing list. Why not move to a web forum while we're at it?

> Sorry to be ranting loudly, but this all seems way too much elitist talk 
> to me.

While I don't agree with some of the reasoning in that document I still
believe in choice, and that requires an educated user.
Yes it would be nice if things would "just work" but I don't like this
"users are too dumb to learn" because that's just insulting.

So what it basically comes down to is "it's a matter of opinion" so this
is basically something the majority can decide.

I vote for "no" reply munging, as it makes *my* life easier and the
privacy of me and others safer.
- --

Dirk-Lüder "Deelkar" Kreie
Bremen - 53.0952°N 8.8652°E

Version: GnuPG v1.4.6 (GNU/Linux)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org


More information about the talk mailing list