Message-Id: <199512292039.PAA12856@jekyll.piermont.com>
To: paulh@imc.org (Paul Hoffman)
cc: drums@cs.utk.edu
Subject: Re: The use of 8-bit 
In-reply-to: Your message of "Fri, 29 Dec 1995 12:33:36 PST."
             <v02140405ad09fe2b1696@[165.227.40.21]> 
Reply-To: perry@piermont.com
Date: Fri, 29 Dec 1995 15:39:11 -0500
From: "Perry E. Metzger" <perry@piermont.com>


Paul Hoffman writes:
> >It seems to me that it is impossible to stop this practice.
> >If it cannot be stopped, perhaps it should be accepted.
> 
> I thought the problem was that these messages may traverse over 7-bit
> networks on their way to the recepient.

Usenet stuff typically travels over eight bit clean paths for most of
its movement. Sometimes this isn't true, but it is true often enough
that large segments of the net (like the entirety of the former Soviet
Union) depend on being able to do things like embed COI-8 Cyrillic in
the messages and have it "just work".

Clearly, we cannot endorse current practice -- different machines
interpret the same bytes differently because some people are expecting
one eight bit character set and others expecting another without any
marker of what the contents actually are. On the other hand, we are
not in a position to dictate, and we cannot actually fix the
problem. I would suggest that we not open up this can of worms unless
someone has a brilliant solution and can get worldwide buy-in on it.

By the way, in general, we are going to face more and more character
set encoding problems in the future. The IETF is going to have to
start worrying about this for real.

Perry