On Wed, 23 Oct 91 13:29:29 -0700, Marshall Rose wrote:
If we agree that a type can have at most one subtype attached, then we agree
that having a single T/V syntax for all parameters is good, and further
that char sets should be part of the parameter list.
Great, let's turn to the next big point.
A requirement was added in the new draft that bodies must have a Content-Type
and a Body-Version header. I see this as a needless requirement. It has the
effect of meaning that every `traditional' message must have the lines:
Body-Version: RFC-XXXX
Content-Type: TEXT; charset=US-ASCII
added to its header.
There is another evil thing about this requirement. I'm surprised the special
interest group for MULTIPART/DIGEST hasn't caught it. It effectively
abolishes MULTIPART/DIGEST. Why? Because it is no longer allowed to default
the Content-Type. The only special thing about MULTIPART/DIGEST is that it
changes the Content-Type default from TEXT to MESSAGE.
This is a glaring inconsistency. True, I will shed no tears for the loss of
MULTIPART/DIGEST. But at the very least the requirement for a Content-Type
must be abolished, and replaced with an explicit statement that the default
Content-Type is always `TEXT; charset=US-ASCII'.