----- Original Message -----
From: "Bruce Lilly" <blilly(_at_)erols(_dot_)com>
On Tue October 12 2004 23:04, Hector Santos wrote:
"Reply to Sender"
You are in danger of confusing yourself (or of being
confused by non-standard terminology used by your
MUA); you probably mean "Author(s)", though it is
conceivable that some respondents might wish to
communicate with a message sender (as indicated
by the Sender field if present, or via the envelope
sender).
No, I think I am pretty verse of the MUA "standard practice" having written
few myself, following consistency in competitive products, and also from the
host side making sure, the support for all of them is consistent following
the standard and BCP. Sure, on a personal level, its so old, that one gets
too confortable to the point I am not even sure that the debate is all
about. I mean, we are trying to get the MUA to do something it was simply
not trained to do. You have Online Readers and Offline readers. Which one
are we concern about? How about those with "exchange-like" modules?
Independent of the actually "UI" terminology used, I am pretty confident the
"standard" RFC based MUA has the following two basic supports:
EMAIL
Reply to Sender -> Offline Direct to Sender
Reply to All -> Offline Direct to "List" of users.
NEWS
Reply to Group -> Online Post to Group
Reply to Sender -> Offline Direct to Sender
Anything beyond this common expectation? That is how we have to make sure
our backends, gateways, expanders must work to support the common
expectation of the RFC-based reader MUA. We are not talking about IMAP or
Exchange or Newsgroup Emulators or are we?
"Reply to All"
and how these two MUA "standard" reply actions are handled or "prepared"
by
list handlers.
They aren't; responses are generated by a respondent,
usually with an MUA -- the mailing list expander is not
involved in that process.
If sent to the list, it list server has a big role here. I explained how
the list server used here seems to not do this, but our list server will
make sure that the proper Reply-To: is used for the distribution. There is
nothing technically wrong with that but it does have an effect on the MUA
for off-list communications. The MUA can be smarter there as you pointed
out.
So to help facilate the MUA, the list handler
should prepare the headers to best provide the most meaningful or
correct
behavior as expected by the user.
Message headers' fields set by authors or whose
semantics are reserved for authors' use should not
be modified by list expanders.
I don't agree when it comes to the network control headers. I think the
backend needs to make that the input is correct and output is correct. Of
course, I am not talking about mail tampering. I am talking about a
groupware concept - mailing list.
The problem as I see it is that we have a standard for 1-1 (direct)
email
messages, but no standard for a list distribution to handle the above
two
reply options.
Incorrect. In fact mailing lists predate email, initial.....
and I am sure there was a time where engraved flintstones were used to
communicate. :-) This doesn't change the fact the current RFC-based mail
readers are what they are today based on an EMAIL and NEWS design. Not
MailIng List concepts. Is this correct or not? In general, groupware is
currently done using something like IMAP, Exchange Interfaces or NewsGroups
Emulation and in a growing large part - a return to online hosting
communications. No? I think so.
....SMTP itself has explicit provisions for
mailing lists.
But the MUA does not. Hence this debate?
In one case (the list server used here), the "Reply to Sender" will go
to
the author. In our list server, the "Reply to Sender" goes to the list
address.
There's that problem with use of non-standard terminology.
I beg to differ. Again, independent of the exact usage of the CUI (Common
User Interface) language used to expose the various reply actions, it is
generally what I outlined above. I don't see how you can see it is not
simply because there is no other technology that the common/standard
RFC-Based reader MUA was trained to work with. That is basically the point
I am making.
In essence, we are trying to deal with a mailing list as it was really a
"newgroup" per se. In fact, I find the handling of a mailing list is more
suitable using a Newsgroup emulator for the reasons of providing a more
natural "reply to group" and "reply to sender" actions. The only think
missing would be a "reply to all."
I guess the difference is that we make sure the "Reply-To:" address is
properly set in the distribution.
Is that more correct?
No. Its use is reserved for the author to indicate where he
wishes responses to be sent.
Hhmmm, sure, for creating the reply address in the new/response message. But
the Target (TO:) is taking from the Reply-To: in the original message.
Again, that is the standard behavior of the MUA. Maybe this where a
Reply-To is changed to MFT before adding the Reply-To: list address?
[more problems based on non-standard terminology elided]
Unfortunately, atleast with Outlook the "Reply to All" does not grab
the "From:" address.
Sure it does -- the problem is that in your case something
has overridden it by setting the Reply-To field.
Ok, good point.
In this regard, I would say our list server makes it more difficult to
send
direct. The user has to understand whats going on and copy and paste
the
From: address from the message properties.
That's a problem with that specific MUA.
I still say there is a conflict either way you look at it. A list expander
adding the Reply-To: list address is more technically correct in my view.
The point is the MUA is not trained to understand a "mailing list."
a) Does a MUA need to detect a Mailing List Message as oppose to a just
a
pure email message?
No, and in general it cannot do so.
Why? Please give a reason. I mean, I would love to put my trust in your
expertise. :-) But I need a reason why this will not "help" in addressing
the issue?
c) What are the most desirable reply/new mail options for a mailing
list?
Reply to List (very similar to a news Reply To Group idea)?
There is no general mechanism to identify a list, messages
may be sent to multiple lists, etc.
I think most List Servers do add some information to pick up on this. I
think List-Address: is a good candidate. No? For multiple list, you would
add the list address(es) in the single header just like its done for
Newsgroups:.
List-Address: address1, address2, etc
Will that work?
Reply to Directly to Author?
Probably yes -- many UAs do provide that functionality.
Reply to All?
Probably yes as a catch-all.
Is that enough?
Probably not; there should be a function to respond to where
the author has requested responses to go. Generally that is
labeled simply "Reply" in UAs.
Yes, but for a direct 1 to 1 EMAIL concept only. That is the issue as I
see. The MUA is going to see the output of a list as "email" and thus
behave as it was a 1 to 1 EMAIL communication. Hence the "reply" as you say,
should go to the requested address. The issue I think is depending on the
list expander, it can go to the Author or it can go to the List. What is
better based on current MUA practice?
How about follow ups?
[...]
Does that make sense?
No. How did you intend "follow up" to differ from "reply"
(which in any case would be more accurately be called
"respond")?
It depends. Follow up has been designed different ways. I mean consider the
design that in some cases only the original author can follow up on his own
mail. :-)
Thanks for your comments Bruce. I hope it can lead to something, rather than
nothing.
---
Hector Santos, CTO
WINSERVER "Wildcat! Interactive Net Server"
http://www.santronics.com