I would like to see a stronger statement that implementations *should*
discard the "prefix" and "postfix" areas. Don't recommend the use of
them. I am afraid that if a sending UA puts an important (judgment
call) message in the prefix area, but the receiving UA discards it.
In my opinion, these 2 UA's are *not* interoperable.
I'm happy to say they should be discarded. But are you also arguing
against the short textual message saying, in effect, "this is a
multipart message; if you're seeing this, you've got a problem"? That
still seems like a good idea to me.
All I am asking for: RFC-XXXX must state clearly that all UA's and
(822<->X.400) gateways should discard them and must not convert these
areas into textual body parts (it is IA5text in X.400.) These areas
should not be exposed to the users.
Content-Label is set by the *sending* UA (e.g. the sender assigns a
label "Phone_Message" to an audio typed body part) and it is more
user-friendlier than Message-ID when used in a reply message. This
field exists in each body part, but it is optional. Message-ID is
different from Content-Label that Message-ID refers to the whole
message and Content-Label refers to a specific body part in a message.
Ah, but in a multipart message, each "body part" is an encapsulated
message, and can therefore have its own message-id header. Therefore
they are, I believe, functionally identical. I'm also not sold by the
argument to "user-friendliness" because I think that, in either case,
the header field is intended for software use rather than for human
reading. Or is this assumption incorrect?
The value in Content-Label is controlled by the user, but Message-ID
is not. The information in Content-Label is intended for human being.