Marshall -
The tspecials problem is a serious one for those who implement the
specification. Unconditional quoting of parameter values, instead of using
tspecials, may help in making sure you never omit necessary quotes. That
isn't the issue. The issue is that if you believe the specification, then you
will not correctly parse MIME that is being commonly transmitted today, or
even some of the examples in the specification!
I am not so sure it is possible to recall all the implementations which
fail to quote parameter values with dots in them. I'm pretty sure that some
versions of MetaMail had this problem, because it was experimented with here
and we encountered the interoperability problem.
The extraneous blank line is more than an annoyance. It too was missed
by some clueless implementors. One of them was me. There is also the problem
of a possible blank line eater at the beginning of message bodies.
A specification that must be violated in order to ensure interoperability
is, in my opinion, broken. So is a specification which suggests to
implementors to do the wrong thing, particularly when those implementors
include Nathaniel and me.
We have the opportunity to fix this problem before it gets worse. I hope
we do.
-- Mark --