[arch-haskell] List behavior (was "Layout of ABS tree....")
xyne at archlinux.ca
Thu Oct 21 21:57:57 EDT 2010
> On the grub-dev list which is also configured with Reply-To, I've
> already twice sent two messages publicly that I meant to be private,
> thus spamming people like you. (Hmm, I'd think my client could notice
> that there's a List-Id or maybe List-Post header and therefore warn me
> if I use "Reply (to sender)" and one of the recipients is the list. Or
> something like that. I'm not sure why Thunderbird doesn't do that.
> Would it violate RFC?)
I don't know what the RFC says but this behavior is logical to me. The original
sender of a message on a mailing list didn't send the message to you (just to
be clear, I'm using general "you" here, not personal "you"). He sent it to the
list and then the list sent it to you and everyone else following it. For each
recipient, the source/sender is the list itself so "reply to" logically sends
the reply to the list, not the original sender.
I really see it as being analogous to partaking in a group discussion. If you
respond to something that someone has just said, it's logical that everyone
should hear you. If you want to reply privately, you should pull the person
aside and do so, i.e. compose a private message. The point of the analogy is
that within the context of a mailing list, the interaction should default to
the list, not private messages. Maybe the relevant RFC says otherwise, but this
is what I consider the most intuitive.
> I mean, if "reply to all" instead said "To:" the list and "Cc:" the
> other people like you, you would be equally annoyed (or almost as
> annoyed, at least -- the philosophical annoyance would be less bad).
> That's the sense in which "to" and "cc" are the same. I already gave up
> on always changing "cc"s to "to"s in e-mail exchanges with the three
> other people in my family, because my email client doesn't help me (and
> I've never heard of one that does), and it doesn't make a technological
> difference .
I understood that and agreed. I only meant that what happens after that is
> That's a good point about filtering into directories though. In
> Thunderbird, I dislike being to/cc'ed because it doesn't put the message
> in the corresponding list's directory; but on the other hand I do like
> being to/cc'ed because then I'm more sure to notice when someone replied
> to me (maybe there's a way to set filters so that I notice all replies
> to my post, even the ones I receive only through the list? I haven't
> figured that out yet.)
This mostly depends on the features of your client. I use claws-mail with
"thread view". It makes it very easy to follow separate threads on the list as
the name implies, including replies to your own posts (and it also makes it easy
to ignore uninteresting threads).
> > Given, this may be a client issue and in general I think people should fix
> > "broken" software rather than expect others to find a workaround for them, but
> > if a single extra header resolves the issue and is commonly employed on related
> > lists then I don't see the problem. Furthermore, arguing that each recipient
> > should implement a manual workaround each time a message is sent instead of a
> > simpler global workaround on the list doesn't make sense. Neither fixes the
> > problem at the source but one is far simpler than the other.
> Does the header resolve the issue, though? On the Arch lists and GRUB
> list I mentioned above, I have been cc:ed in responses to some of my
> posts. I think this is because many people without "reply-to-list"
> functionality have no choice but to get in the habit of using "reply to
> all" to reply publicly to a mailing-list post (unless they only live on
> mailing-lists that set Reply-To).
I don't know as I don't use a (noticeably) broken email client. :P
I can only say that with claws-mail, it seems to work. All I have to do is hit
"reply" and my message goes to the list. It seems to work for everyone else
too as I don't think I've ever been CC'ed by others replying to the list.
More information about the arch-haskell