Keith Moore <moore(_at_)cs(_dot_)utk(_dot_)edu> writes:
- to me, whether you want personal copies on a reply to a list seems
like a sender preference rather than a list preference. it
depends on how much mail you get, how you process mail, etc.
Although one thing that has to factor into this is whether the person
is on the mailing list at all. I always want personal copies of mail
to lists that I'm not on, for instance.
for me, it's obvious that I do want separate copies, because I
have all of my list mail sent to sub-folders, and I like to
arrange things so that my inbox has a history of all of my
correspondence and the replies to my correspondence, while my list
folders are complete archives of the list. but I can certainly
imagine how someone with a single mail folder would not want to
Of course, part of the problem is that it's obvious to you, but not
obvious to anyone corresponding with you.
Nor is it obvious to someone corresponding with me whether I'm
subscribed to the list. (For instance, I post to the ietf list once in
awhile, and I occasionally even "reply" to messages, but I'm not
subscribed any more. I scan the archives every few days.)
One could jump to the
conclusion that you do because you send them, but that frequently
isn't the case; there are people who send them and don't want them
I'm not sure whether you're right that personal copies are always a
sender preference. You may be. I can't think of a good
counter-argument right now. (Whether replies should go *only* to the
individual or also to the list is often a list preference, however.)
There are certainly lists for which it is never appropriate to copy them
on replies. For instance, there are announcement-only lists.
Incidentally, it's not just a question of single folder vs. multiple
Agree. I am just citing that as a convenient example.
(aside: I keep wanting an MUA that will give me views of messages
independently of folders. For instance, I want a view that contains
every message in every thread I've contributed to, even though those
messages might actually be in a variety of folders. If I had that
kind of MUA, I might be quite happy to not be cc'ed on replies that
included any list I happened to be subscribed to, because my MUA
would find the message in the folder for that list. OTOH, such an
MUA would certainly need to be able to detect duplicates anyway.)
- a separate issue is whether an address continues to get copied
on subsequent replies. (A sends a message to the list, B
replies with a message To A and CC the list, C replies to B's
message, CC'ing A and the list. A gets CC'ed on subsequent
messages in that thread). *that* might well be a list preference.
Good point. It would indeed be better to deal with that as a separate
issue, since that's where most of the complaints come from that I see.
The quantity of mail is larger if Cc's keep being preserved, so while
people are willing to delete one message, they get annoyed by deleting
One wrinkle might be that this is sometimes related to topic drift.
some folks are fine with the first few cc's but as the discussion
changes focus those people are less relevant. I don't have any good
ideas for how to deal with that in the context of email discussion
groups - for a discussion medium to deal with that problem reliably and
effectively I suspect it needs a very different architecture than
either email or usenet.
- different communities do have different ideas on the best way
to handle these issues. often this is a reflection of
the technical sophistication of each community and the kinds of
tools they tend to use. at other times it's a reflection of
the kinds of conversations those communities want to have.
In my experience it really is more the latter than the former. Much
is made in this debate of how one side or the other believes what they
do because they don't understand how mail works, but that hasn't been
My experience is that users with different levels of sophistication
will want to solve the same problem in different ways. A more
sophisticated group will be more likely to file incoming messages
into folders, set up filters to cull out messages they don't want
to see, etc. - because they want more flexibility out of email.
A less sophisticated group will want reply-to to point to the list,
subject fields to be munged to include the list name, etc. They
may or not be aware of the problems caused by these things (some
are, some aren't) but in either they don't care about the problems
because they don't miss the functionality they're losing.
I don't necessarily think it's useful or appropriate for technical
standards to accommodate everybody's different ways of dealing
with a problem. I think the technical standards need to define a
small number of ways to deal with the problem (ideally one way)
and try to make sure that that way will work for the diverse sets
of users out there.
That being said, I think any solution to the list reply problem that
doesn't deal with list preference for personal replies, list
preference for replies back to the list, personal preference for
copies of list replies, and personal preference for no copies of list
replies, plus still copying people who aren't on the list, is
insufficient to the task.
Hmm. It might be useful to try to design a system with those
parameters and see what it looks like. I'm thinking that if an author
says he wants a personal reply, he should definitely get one - but if
the list says that the author should not get a personal reply, maybe
the personal replies should be Bcc'ed (or Cc-No-Reply'ed). Or to put it
another way, maybe the list's preference is properly not whether
personal replies should be sent, but whether personal replies should be
visible to the list.