MFT can cause the replies to go to some other set of recipients
without the responder realizing this.
My MUA shows me the original message along with the reply under
construction. To-line differences are already reasonably clear, although
I suppose automatic highlighting of the differences in red would save a
moment of reading.
But the point you keep missing is that the original sender and the
original recipient are _cooperating_ to send followups to the right
place. Sure, there are occasional disagreements, but _usually_ the
recipient wants exactly what the sender specified in Mail-Followup-To.
The way _you_ use MFT, this is _usually_ what the recipient wants.
It's not clear whether this would be true if MFT were more widely
adopted by a more diverse set of users.
But I do agree that the user interface can make a huge contribution to
helping the responder be aware of what he is doing.
USENET has always allowed the original sender to specify Followup-To.
The recipient can select a different followup target but _usually_ wants
exactly what the sender wants. That's why the recipient's software uses
Followup-To as the _default_ followup address.
USENET is a different environment, with a different purpose than email.
Many assumptions that are reasonable in USENET are less reasonable in
email. For instance, it's meaningless to talk about USENET discussions
that don't involve any newsgroups, but it's perfectly reasonable to have
email discussions that don't involve mailing lists.
Keith