I've been thinking about the taxonomy. I'm wondering if it would be more
useful to describe automated messages in terms of:
I don't see how your proposed taxonomy helps. as far as I can tell,
sender expectations (e.g. did the sender set reply-to anticipating
that a human would respond?) and whether the responder configuration
is under control of the recipient (i.e. who is responsible for
maintaining this responder?) should govern the responder's behavior,
not the nature of the content in the response.
Keith