🚀 go-pugleaf

In article <rhoward.726332201@matd>, rhoward@matd.gatech.edu (Robert L. Howard) writes:
> What happened to the discussion on using MIME to handle a MUA
> level Return-Receipt (RR) mechanism?
> 
> Some points:
> 
> 1.  It MUST be voluntary.  Ideal behavior would be for a MIME
>     MUA to ask, "Sender has requested return-receipt, would you
>     like to return-receipt on this message? (y/n)".  MUAs
>     could have user configuration parameters to a) always send
>     RR without asking b) always ask or c) never ask and never 
>     send RRs.
> 
> 2.  I don't particulary care how this is to be handled MIME header-
>     wise (could someone recall previous discussions).
> 
> 3.  There should be a unique ID string type thing for RR message
>     parts so that when the receipt comes back the MUA can easily
>     match it to the message the receipt is for.

I don't see any problem with it, so long as it's clear that what's
being provided is an (optional) "read" acknowledgement, not a
"delivery" acknowledgement.  Basically someone needs to define a MIME
content type for read acknowledgement requests and another type for
read acknowledgements, and define how a supporting mail reader should
act when it receives a message containing a read acknowledgement
request.  Also needed is a definition of exactly what a read
acknowledgement means.

Note that the mail reader, having generated a read acknowledgement,
should probably mark the message in some way, so that duplicate read
acknowledgements are never generated for a message.

The term "return receipt" seems ambiguous.  It seems like some of the
past flaming (thought not all of it) has been due to confusion about
whether delivery acks or read acks were being discussed.

Keith Moore