From: Per Abrahamsen (abraham@dina.kvl.dk)
Date: Tue Jan 08 2002 - 11:06:23 CST
list-ietf-wg-apps-usefor@faerber.muc.de (Claus Färber) writes:
> . 90% can't post to the group at all (most of the time), 30% can't
> even read it but 10% see the name correctly.
> (NB: The numbers < 100% are arbitrary guesses.)
[ In a transitional period for those who cares. ]
> . 100% can read rthe group and post to it but don't see the name
> correctly.
[ Forever. ]
I believe #1 is preferable, the most popular newsreaders already
support UTF-8 in one form or another, so being standard comforming is
cheap. We are already most of the way. Usenet is not WWW or Mail, we
cannot assume unlimited development ressources to implement yet
another character encoding. We have to piggy-back on other standards.
> Languages based on Latin scripts tend to be writeable with the
> basic Latin (ASCII) alphabet. Swedish is an exception AFAIR.
Together with Danish and Norwigian and Icelandic and Finnish at least.
I guess that the Scandinavian regional hierarchies and users will be
the first to utilize non-ASCII group names, in whatever form, at least
among the Latin based language groups. The other non-English Latin
based hierarchies will probably only follow when the names "look
right" on most software. The English based hierarchies when
everybodey has forgotten about the ASCII restriction.
So it will mostly be us Scandinavians who will have to petition the
newsreader authors to support non-ASCII newsgroup names. I believe we
will have a much more convincing if the required changes are
relatively small.
> This is plain wrong. MIME solves the problem of getting 8bit
> characters and binary files through a 7bit mail infrastructure.
Which gave huge amounts of problems, compared to simply upgrading to a
8bit infrastructure. Something that was mostly complete, at least for
the people who cared.
Maybe we Scandinavians felt it more than you Germans for the reasons
stated above, which could explain why both Erland and me are against
inventing new encodings.
> It also solves the problem of exchanging charset and file type
> information.
Yep. The "Content-Type: " header was badly needed. At the time
various incompatible vendor-specific ad-solutions were already
becomming common.
> And it solves problems with preceding standards such as
> UUENCODE (no file type information, no clear labelling, etc., some
> characters are not safe with some mail paths).
Inventing BASE64 instead of reusing a strict form for UUENCODE was
another choice that caused much more problems than it solved. The
effort to upgrade the few unsafe mail paths would be much smaller than
implementing BASE64 support for all mail and news agents. And even
now, UUENCODE seems to be the prefered file transport format on many
(most?) binary groups on Usenet.
> It is debatable if the problem really existed, i.e. if mail really
> was not 8bit clean (it's certainly not binary-clean). The charset
> labelling was a real problem, though, because Unicode was not yet
> invented back then (RFC 1341/2, June 1992).
Agreed.
> Which version? There were reports on this mailing lists, that some
> versions don't work.
I don't think we ever got a followup from the XNews user who reported
problems.