Mark Crispin writes:
I've never heard this before. If it is true, I think the people who came up
with this idea should get a stern talking-to for being so silly. This idea
is completely unworkable in practice.
Ned, I'm sorry, but I must disagree. This is not only workable, it's used!
One of the biggest long-term complaints in DEC-20 MM is that it doesn't
support this -- a mistake I did *not* repeated in my more modern mail
software.
It may be used, but it is still unworkable in practice. This is not a
contradiction. You simply have to clarify what sort of practice you're talking
about. By practice I mean something that extends to the RFC-compliant mailer
software in use on the Internet today.
There is nothing that says I have to list any of the envelope recipient
addresses in the headers, and there are many cases where this is in fact NOT
done. Thus systems that rely on such information being present are therefore
unworkable in practice, since you cannot rely on it in general!
I know that this is true -- I see examples of it every day. My mailbox is
crammed with examples of it every day, as a matter of fact. You have to either
grant me this or else make the claim that usage where envelope recipients are
not represented is incompliant. And that is going to be very difficult to do.
This then restricts your assertion to the notion that it works to the cases
where it does happen to work -- sometimes, maybe, possibly, and on odd days of
the month when the moon is full. This position is one I can accept. I just
don't design mail systems that depend on it. I do design mail systems that
don't strip such information or otherwise interfere with it. Apparently MM did
interfere with it somehow. News to me, but you'd know this better than I.
More important, the use of mnemonic does not impinge on this usage in any way.
Far from it, it actually enhances it (see below).
My earlier statement were based on the proposal to standardize such practice.
Not only is such an act incompatible with the standards as they are now
written, it is simply a terrible idea in general!
I think your problem is that you don't understand that the routing involved is
non-electronic. It's human.
I don't have a problem with this, and I do understand what sort of routing is
involved. I gave a specific example of this in my previous posting!
But human routing is the most flexible in the world. It is the kind of routing
that is most likely to be well-served by mnemonic encoding (this should be
understood from the definition of the word). Mnemonic gives us the best of both
worlds -- phrases humans can understand, no breakage of existing software, and
underneath it provides an infrastructure that stupid machines can deal with
too.
Thus, I'd think that if you want to support this kind of routing while
simultaneously supporting 7-bit-only user agents, you'd be firmly behind the
use of mnemonic, which I see as the only practical means to continue this usage
short of disallowing the use of extended character sets completely. And if
you're in favor of this last, you can then explain to me what the difference is
between that and my using mnemonic without telling you what I'm using...
Ned