Re: the most obvious failure in To-NoReply

[Simon Josefsson <jas@xxxxxxxxxxx>]

Keith Moore <moore@xxxxxxxxxx> writes:

> D. J. Bernstein wrote:
>> Keith Moore writes:
>> 
>>>you're free to set Reply-To to ietf-822@xxxxxxx
>> I don't want _replies_ to go to ietf-822@xxxxxxxx I want _followups_
>> to go to ietf-822@xxxxxxxx
>
> "followup" is a Usenet term.  Usenet has two different kinds of 
> addresses - email addresses and newsgroups.  Responses to the author's 
> email address are replies; responses to newsgroups are followups.  Email 
> does not have visibly different kinds of addresses and does make a 
> distinction between replies and followups.  Email UAs have "reply" and 
> "reply to all" or similar.  For several reasons including both protocol 
> differences and cultural differences between email and Usenet, email's 
> "reply to all" and Usenet's "followup" are not quite the same thing. 

What's the difference?  I view them as conceptually the same thing.

> And I don't see evidence of any desire among users to make email more 
> like Usenet.  (actually I see plenty of evidence to the contrary.  but I 
> digress.)

Perhaps the users of gmane.org would disagree.  I'm reading this
mailing list via news through their services, for example.

>> An increasing number of MUAs support the sender specifying
>> a followup list (normally with Mail-Followup-To) that doesn't include
>> the reply list; that's what users want.
>
> A few MUAs support MFT.  The ones I've seen that support MFT represent 
> an insignificant part of the installed base, so I don't accept an 
> argument that the level of MUA spport for MFT is an indication of user 
> preference.
>
> For that matter, the most widely used OS by far is rife with security 
> holes and has been for nearly 10 years now, but I don't accept that as 
> an indication of user preference for security holes.
>
> It seems to me that you've had a lot of time to try to get MFT widely 
> accepted in the marketplace.  Your failure to do so doesn't necessarily 
> mean that MFT is bad, and it says nothing about the relative merits of 
> MFT vs. NR.  But it probably does mean that claims that equate market 
> acceptance with rightness are moot in this discussion.

Right.  But as far as standardization goes, having a proposal
implemented and tested in practice is a strong argument.  Running
code, and all that.  The problems with MFT appear to be well
understood, and some communities still chose to use MFT.  So I
believe, regardless of any NR vs MFT discussion, that the needs of the
people using MFT should be taken care of by any proposal.

Thanks,
Simon