I think that Nathaniel's observation that there really IS a difference
between a text body part that ends in a newline and one that
doesn't, and that displaying the next body part in a different possition
is an appropriate action is worth taking seriously.
First, there certainly is a difference when the originator chooses to
create a newline, versus when they don't.
Second, Mime has minital rules concerning display choices to be used with
respect to the relationship of body-part sequences.
Third, the receiving UA has to make SOME decision.
Fourth, presence/absence of an ending newline is not entirely under
the control of the originator, given the truly creative behaviors Internet
mail relays and gateways engage in. So, while the receiving UA has to make
SOME sort of decision and the one Nathaniel describes is certainly
reasonable, I hope none of view this issue as a done deal.
Fifth, my comment had not been intended to prescribe or proscribe the actions
Nathaniel discussed. I was meaning to address the question of how the
receiving UA interprets "larger" aspects to the message structure. An
example would be to assume that the presence of the ending
newline means that the next part in the multipart/mixed set is truly
an independent object, whereas the absence of the newline would mean that
the next part is fully integrated with the text part. (For example, an
o/s with a rich file store might store the two parts in two files if
the newline was present and store them in one if it is not.) This is
the sort of overinterpretation of the semantics that I meant.
In truth, it is perhaps appropriate to attend to the core question of
a presentation syntax, to correlate body parts specified by the Mime
structure syntax.
Dave