On Sun, 10 Oct 2004, John R Levine wrote:
> > As has been pointed out about a dozen times in the past week,
> > quite a lot of mail is received by programs that are not MUAs.
> > They include mailing list managers, service gateways, and other
> > stuff. A lot of those programs don't handle MIME at all. If
> > MASS adds MIME sections to mail that used to be plain text, a
> > lot of those programs will fall over and die when they get
> > MASS-ized messages.
> They already handle MIME.
Some do, some don't. As I said in other messages, I know of enough
mail applications that don't handle MIME that I am not comfortable
writing them all off. I also don't know how robust the applications
that do handle MIME are, and how many will break if presented with
MIME sections they weren't anticipating. This is why we need
experiments.
This issue of backwards compatibility has the character of dogma.
John's right, of course, we need experiments.
The question, in my mind, is not whether backwards compatibility is
important but whether backwards compatibility is more important than the
infrastructure we're putting in place. And I don't claim to know the
right answer to that question, nor do I think *the* right answer exists.
It's a judgement call that changes given the environment of the time, in
my opinion.
MIME was relatively unique in that backwards compatibility was critical:
between ordinary text-based email and MIME-based text email as well as
the need to work over a 7bit transport.
Today, most MUAs support MIME, a lot of applications are MIME-aware (to
the extent they process enough MIME to continue to do their job), a few
MTAs support MIME directly (many indirectly in order to provide virus
scanning), and 8bit transport is widely available.
My observation is simply that MIME is quite valuable. It's deployment
to date is evidence of that. It may not be perfect but it's what we
have. It offers a lot flexibility and opportunity for future services
or needs. If we can motivate the integration of MIME into MTAs and
further upgrade or improve the robustness of applications with a
*minimum* of disruption as opposed to no disruption, that's a win. Of
course, agreeing on what "minimum" is becomes the sticking point, the
"judgement".
Somebody (sorry I forget who) mentioned that a "net win" was important.
Well, the "net win" is a reduction in forged email. Since the volume of
forged email is more likely to get worse than it is to get better, that
will provide the motivation needed to deploy MASS.
A sub-observation is that MTAs are going to have to make changes to
support MASS regardless. Adding MIME to that requirement is a
no-brainer in my opinion. The real question is how bad is it to the
rest of the message handling system, balanced against how good it is for
the entire message handling system (keeping in mind it will be deployed
as the need to reduce forged email increases).
Jim