Re: A proposal

[Erik Naggum <erik@xxxxxxxxxxxx>]

> Excuse me (while I kiss the sky), giving only suppport for 8859-1
> would be sub-standard, and not meet the present requirements.

I'm tired of this line of argument.  My reaction to the two year delay
of ANSI C due to some incredible "internationalization" kluge was part
disgust, part curiosity in what makes people do that kind of thing.
Rather than get a good solution with growth potential out so we can
solve 90% of the problem, some people with an overly active interest in
the remaining 10% can stall the whole thing for years.

I'm sure you, Keld, in particular, are aware of the general paradigm of
ISO character sets with the C0, C1, G0, G1 areas.  I'm sure you are
aware of the direction of ISO DIS 10646 with respect to honoring that.
I'm also quite certain that you know that ISO DIS 10646 has ISO 8859-1
in Group 032, Plan 032, Row 032, and that this is the default until
something else is said.  I'm further quite convinced that you are able
to see the similarity between this and using ISO 8859-1 only, in fact,
that the similarity borders on identity.

Thus, for a very basic interoperability when either party cannot handle
full ISO 10646 (which nobody are required to do, yet), it is impossible
by all practical means to distinguish between ISO 8859-1 and ISO 10646
texts.  Don't you think this is fortunate?

Don't you think that we can decide that we allow 8-bit data paths, AND
use ISO 8859-1 as the character set of choice, UNTIL the ISO folks have
quibbled enough over the political conundrums to be able to raise what
sounds like one voice to the uninitiated and declare with varying force
that ISO 10646 is The Answer to the Ultimate Question of Life and the
Universal Code for Everything?

Don't you think that IETF SMTP should NOT aim at solving world problems
of this vast political nature when some other folks are doing the
political acrobatics necessary?  Do YOU want IETF SMTP to ponder the
intracacies of Han Unification?  I hope you don't.

> There is about 2000 sites on EUnet (mostly uucp based).

With all due respect for UUCP sites, we are not discussing UUCP
problems.  UUCP is _already_ 8-bit wide.  I, and my little network of a
dozen sites, use ISO 8859-1 for Norwegian mail.  We use UUCP.  I like
that a lot.  I'd like to make a small step for mankind and a giant leap
forward for mail at the University of Oslo and with my clients, by using
ISO 8859-1.  (Others on this list more properly represent the U of
Oslo, I just work with them.)

> As such, to provide a sufficient service to all our customers,
> we need to be able to handle full Latin (including East-European),
> Cyrillic and Greek.

Great!  I'd love to have a full and sufficient service for all the
customers everywhere, but we CAN'T DO THAT IN ONE JUMP.  There are more
pressing concerns than Cyrillic mail at the moment.  I mean, despite
your frantic concern for the rest of the world, for which they are
probably going to be eternally grateful, you must see that if you stall
this, we won't have "user requirements" satisfied for years to come.  I
assume that "user requirements" are your concern, not something else.
A conspiracy theory where you were to introduce intractable problems to
the Internet world on behalf of the X.400 world fits the data points,
almost as well.

> Wearing my RARE hat, the RARE ad hoc WG on character sets has passed
> a resolution that support for a lot of character sets were
> required, to accomodate user requirements. The character sets
> included IBM and other popular vendors' character sets, and
> also Greek and Cyrillic. RARE is also a major player on the European
> Internet scene.

WHAT!?  You required support for _IBM_ character sets "and other popular
vendors' character sets", and still you have the chutzpah to talk down
the efforts to use ISO 8859-1 in Internet SMTP mail?  Get out of here!

Let's resolve to attack and solve real problems, folks.  This is not
one of them.

And now for a "claimer":

I'm Norwegian.  I'm involved with ISO 10646 on an informal mailing list
basis.  I've contributed to ISO 6937-2:199x for Norwegian.  I've tried
to reconcile SGML entities with ISO 10646 symbols.  I think character
set issues are terribly important.  I know that we not only open a can
of worms, but a barrel of same, if we try to take on that issue, also.
I'm getting involved with ISO/IEC JTC1/SC2, SC18, and SC22 for Norway,
but it's a slow process.  I'm moving in on POSIX and ANSI C to design
proper locales for this country (Norway).  This issue is not one we
_want_ to tackle in this working group.  Let's get liaisons, and let's
stay out of the other people's agendas, OK?

[Erik Naggum], pissed at all the non-issues and the noise they produce.