Keld writes, referring to a message of mine...
I would expect the above agreements might be very hard to get in a
network context that hasn't been able to agree on how to map 7bit ASCII,
but that isn't a MIME problem.
Much of this is done in RFC1345: definition of a number of charsets,
including many EBCDICs, reversible mapping rules and fallback notation.
Keld,
I think there are really three separate problems here, and thinking
about them separately is the only way to move forward.
(1) An adequate inventory of the various options, including the relevant
EBCDIC and ISO character sets and possible mappings. RFC1345
accomplishes that at the 95% level, if not at the 100% one.
(2) Agreement (in this case) within and between the CREN, EARN,
NetNorth, GulfNet, etc., communities about which EBCDIC(s) to use, which
translation table(s) to use, and how to document what is being used.
Note that each of these has its own decision-making arrangements and its
own mechanisms for enforcing its decisions (or the absence of these
things). Failure to be able to make network-wide decisions in this area
has been a problem long before either MIME or RFC1345 existed--the
arguments and problems are at least a decade old.
(3) Agreement about how to handle MIME and its charset parameter in the
above context.
In addition, if whatever solutions are adopted require any sort of
deployment at all, there is a deployment problem in a connection of
networks have hosts that have proven, in the past, very resistant to
uniform deployment of almost anything that has been thus attempted.
It is my personal belief that the total confusion in the universe is
increased by believing that the second and third problems can be solved
simultaneously. But perhaps I'm wrong. But no solution to one of the
problems constitutes a solution to the others.
john