headers
Bruce Lilly | 1 Sep 01:41
Picon

Re: [Fwd: I-D ACTION:draft-moore-mail-nr-fields-00.txt]


Keith Moore wrote:

> I
> suppose you could argue for a UA that in the presence of list-post, sent
> replies to that address instead - but I think that would be very
> counterintuitive.

I know of at least two MUAs that provide three reply options:
    Kmail 1.7          Ximian Evolution 1.4.6
---------------------------------------------
Reply to Author           Reply to Sender
Reply to All              Reply to All
Reply to Mailing List     Reply to List

In each case IMAP folders can be configured as associated with
a mailing list mailto URI which is used for the Reply to
[Mailing] List function.  No List-Post field required.

Both are of course limited somewhat in terms of platform support.
Permalink | Reply | 134 comments |
headers
Claus Färber | 1 Sep 14:58
Picon

Re: [Fwd: I-D ACTION:draft-moore-mail-nr-fields-00.txt]


Simon Josefsson <jas <at> extundo.com> schrieb/wrote:
> This might be, but at least my experience with deployed MUAs suggest
> this isn't what is implemented today.  MUAs prefer the Reply-To over
> From on "reply author".

Well, the "reply author" function is IMO just mislabelled. It actually
means: "reply to a small number of persons".

Most MUAs (the M here is for "message", including Netnews) have two
pre-defined reply functions: a narrow reply going to a small number of
persons and a broad reply going to a large number of persons, often via
newsgroups or mailing lits.

These can be called narrow/broad, reply/followup, reply-to-author/reply-
to-group, private/public, etc. but they are basically the same concept.

Strictly, mail does not know that concept but MUAs try to provide it to  
the users. Mail-Followup-To provides that; this is why it's popular with

Keith Moore's proposal caters for the same problem (replies don't go  
where they are supposed to go) and it's more compatible with existing  
MUAs that don't understand the header. But it's more complex to under- 
stand for the average user and it's not compatible with the mental model  
users have made based on the common user agent UIs. And it still does  
not provide a distinction between different types of replies, which is  
what user agent desingers want to provide.

Claus
--

--
(Continue reading)

Permalink | Reply | 2 comments |
headers
D. J. Bernstein | 1 Sep 21:10
Picon

Re: reply problem list


Bruce Lilly writes:
> If the original message author wishes responses to go
> to the list, he should set Reply-To to the list address.

Senders often want to _encourage_ followups and _discourage_ replies.
That doesn't mean senders want to _break_ the reply mechanism.

A new header field could formally declare the same information:

   Are-You-Sure-You-Want-To-Reply: I'd like to see this topic discussed
     publicly.

An MUA that sees this header field could warn the replying user at the
beginning of composing a reply:

   Are you sure you want to reply privately? The message author says:
   "I'd like to see this topic discussed publicly." Click here for
   private, here for public, here to cancel.

This approach doesn't screw up the Reply-To information, so the user can
still send a reply rather than a followup if he wants to.

---D. J. Bernstein, Associate Professor, Department of Mathematics,
Statistics, and Computer Science, University of Illinois at Chicago
Permalink | Reply | 9 comments |
headers
Keith Moore | 1 Sep 21:57
Picon

Re: [Fwd: I-D ACTION:draft-moore-mail-nr-fields-00.txt]


> 
> Simon Josefsson <jas <at> extundo.com> schrieb/wrote:
> > This might be, but at least my experience with deployed MUAs suggest
> > this isn't what is implemented today.  MUAs prefer the Reply-To over
> > From on "reply author".
> 
> Well, the "reply author" function is IMO just mislabelled. It actually
> means: "reply to a small number of persons".

Or to put it another way: MUAs really need a "reply author" button,
but they don't have one.  The typical "reply" button doesn't
quite work like it should because the "narrow reply" really doesn't
map to what the recipient is likely to want to do.

(The only time I ever use (narrow) "reply" is when I really want 
"reply to author".  And then I have to check the From field of the
resulting message to see if the author had specified reply-to .
If he did I often have to edit the field and restore the correct
address.)

> Keith Moore's proposal caters for the same problem (replies don't go  
> where they are supposed to go) and it's more compatible with existing  
> MUAs that don't understand the header. But it's more complex to under- 
> stand for the average user and it's not compatible with the mental model  
> users have made based on the common user agent UIs. 

Actually I think it's the other way around.  Setting mail-reply-to
and mail-followup-to headers is more complex for the average user,
since it requires the user to specify exactly which recipients should
(Continue reading)

Permalink | Reply | 1 comment |
headers
Keith Moore | 1 Sep 22:07
Picon

Re: reply problem list


> 
> Bruce Lilly writes:
> > If the original message author wishes responses to go
> > to the list, he should set Reply-To to the list address.
> 
> Senders often want to _encourage_ followups and _discourage_ replies.
> That doesn't mean senders want to _break_ the reply mechanism.

The reply mechanism is already broken.  The way Reply-to is implemented,
it doesn't do the right thing for either "reply to all" or "reply to
author".

There are lots of ways to fix this, but they all involve changing 
how user agents work to some degree.  To people who are adapted
to the current broken behavior, those changes will be seen as new
broken behavior - at least until they get used to it.  The real
question is which approach produces the smoothest transition and
the most favorable end-state.

Keith
Permalink | Reply | 0 comments |
headers
Tony Hansen | 1 Sep 22:14
Picon
Favicon

3rd party ietf-compliance testers


Do anyone know of any companies out there that will do IETF compliance 
checks of SMTP and/or POP servers? We have a situation where we could 
use such a service.

Thanks!

	Tony Hansen
	tony <at> att.com
Permalink | Reply | 1 comment |
headers
Bruce Lilly | 2 Sep 01:45
Picon

Re: reply problem list


D. J. Bernstein wrote:
> Bruce Lilly writes:
> 
>>If the original message author wishes responses to go
>>to the list, he should set Reply-To to the list address.
> 
> 
> Senders often want to _encourage_ followups and _discourage_ replies.
> That doesn't mean senders want to _break_ the reply mechanism.

Mail doesn't have "followups", it has replies.  The Reply-To field
provides a sender with a mechanism to specify the reply address(es)
which he wishes to encourage.  It is of course still possible for a
recipient to direct replies elsewhere.
Permalink | Reply | 7 comments |
headers
Simon Josefsson | 2 Sep 02:35

Re: reply problem list


Bruce Lilly <blilly <at> erols.com> writes:

> D. J. Bernstein wrote:
>> Bruce Lilly writes:
>> 
>>>If the original message author wishes responses to go
>>>to the list, he should set Reply-To to the list address.
>> 
>> 
>> Senders often want to _encourage_ followups and _discourage_ replies.
>> That doesn't mean senders want to _break_ the reply mechanism.
>
> Mail doesn't have "followups", it has replies.  The Reply-To field
> provides a sender with a mechanism to specify the reply address(es)
> which he wishes to encourage.  It is of course still possible for a
> recipient to direct replies elsewhere.

Real world mail has more than one kind of reply, call the distinction
public/private, narrow/broad, reply-author/reply-all or what you will.

Setting Reply-To might work for one kind of replies, but does the
wrong thing for the other kind of replies.  To see this, just go
through, for each kind of reply, what happen if Reply-To is set to the
mailing list address or to nobody:;.  This is why we need a better
reply mechanism for mail.

The hope should be that a better solution than forcing users to do
manual work can be worked out.
(Continue reading)

Permalink | Reply | 1 comment |
headers
Bruce Lilly | 2 Sep 03:08
Picon

Re: reply problem list


Simon Josefsson wrote:

> Setting Reply-To might work for one kind of replies, but does the
> wrong thing for the other kind of replies.  To see this, just go
> through, for each kind of reply, what happen if Reply-To is set to the
> mailing list address or to nobody:;.

That will also depend on what is in To and Cc fields of the original.

For a partial analysis see http://www.landfield.com/usefor/2004/Jul/0274.html
Permalink | Reply | 0 comments |
headers
Charles Lindsey | 1 Sep 21:40
Picon
Picon

Re: [Fwd: I-D ACTION:draft-moore-mail-nr-fields-00.txt]


In <284CE4D6-FB84-11D8-B65D-000393DB5366 <at> cs.utk.edu> Keith Moore <moore <at> cs.utk.edu> writes:

>there is no solution to these problems that doesn't involve some 
>changes in MUA behavior.  the question is:  which changes give the most 
>benefit for the least transition pain?

I think the first thing you need is a third way of replying -
"Reply-To-List" which, as Bruce had told us, some MUAs already provide.

And then you need some facility for an original user to say whether he
likes or abhors personal replies in addition to replies to the list.

--

-- 
Charles H. Lindsey ---------At Home, doing my own thing------------------------
Tel: +44 161 436 6131 Fax: +44 161 436 6133   Web: http://www.cs.man.ac.uk/~chl
Email: chl <at> clerew.man.ac.uk      Snail: 5 Clerewood Ave, CHEADLE, SK8 3JU, U.K.
PGP: 2C15F1A9      Fingerprint: 73 6D C2 51 93 A0 01 E7 65 E8 64 7E 14 A4 AB A5
Permalink | Reply | 52 comments |

Gmane