Is it reasonable?

[Andr'e PIRARD <PIRARD@xxxxxxxxxxxxx>]

I've just been confirmed that SunOS 4.1 does implement 8-bit mail with
ISO 8859-1 (those fonts are in the system). Gracefully. End of quote.
So, I can see that "foreign" computers also realize that they've got no
less bits in their bytes than ours, if you don't mind my way to put it.
And that they find great that the communication line "happens" to have
exactly as many. Will we be able to stop this trend and find an excuse to
have left that forbidden sharp knife on the table corner?

I appreciate the concern to guarantee 8-bit transport.
But does *that* *require* encoding?
Am I wrong thinking that most Internet mail is end to end, that the only
MTAs requiring a trivial fix are Internet *relays* (as opposed to end ones),
that gateways to other transport protocols is just half SMTP's matter but also
half SMTP's courtesy?
Is St Louis' conclusion really that each existing or maybe easy to convert
8-bit UA will require end MTA extensive modification in addition?
My main gateway is my own Internet/BITNET one, written by IBM, same as in CUNY.
Are you sure you can convince IBM people to implement 7 bits encoding?
*Their* wink is not just amused, *they* do 8-bit on *their* networks.
I even see that their various SMTP's are 8-bit wide, but without concern for
useful transcoding nor concept of a standard code on the line.
I'll be able to 8mail across the lawn to the aforementioned Suns in some time.
Will they have to be convinced too that the only working thing is a bug?

Even if I feel that 8-bit mail is no more system-blasting than mailing control
characters or making ISO 8859 file FTP-available, I respect this concern.
The European attitude is by no means aggressive or whatever, but understand
that one side of the globe cannot walk on hands to match the other's position.
I mean neither side, but here, 7 bit restrictions are spoken of with a wink.
Let's concentrate on finding how 8-bit traffic cannot intrude no-thanks
systems and let the migration be progressive in the right direction.
The simpler the better. The best machines are those with the fewer parts.

If I could restrict dns access by our mailers to those accepting 8-bits and
relay the rest to a stripper, I'd do that gladly!!!

In fact that's only the easy part of the story.
I've been so long and deep involved in international character sets problems
that I have to stress again how lucky you are with an unquestioned ASCII code.
Believe me, not defining what is the one and only code to be used in the
*whole* TCP/IP protocols suite is a careless attitude leading to chaos.
What code is used inside a system is their matter. What code they put on
the line is everybody's, unless it can be *certain* that the other side is
exactly of the same system species. And even then, this leads to user's
confusion.
X have understood this apparently.
tn3270 completely missed the point and I have spent hours explaining EBCDIC
to ASCII victims.

Robert Ullmann's 8-bit draft is the metric system if ISO 10646 will be too.
While waiting, ISO 8859-n can be tolerated between like systems as a
temporary measure between like systems.

Keep showing the way!

Andr'e PIRARD             SEGI, Univ. de Li`ege
B26 - Sart Tilman         B-4000 Li`ege 1 (Belgium)
pirard@vm1.ulg.ac.be  or  PIRARD%BLIULG11.BITNET@CUNYVM.CUNY.EDU

PS: I hope I'll have time to rework some text I want to make FTP available.
All of ISO 8859-1 it will contain is pictures ;-)