My main concern thru this discussion on news and also the MIME
discussion some years ago, is the interoperability between the
new standard, and the old way things were done.
We see from MIME experience that the transition period will
be measured in years. It is now more than 2 years ago that the MIME
rfcs appeared in their first issue, and although MIME is spreading,
I see people turning back to other mail formats, after trying it,
and then being flamed for it. So MIME spreads even more slowly
than normal upgrading habits would expect.
This reluctance appears regulary in my environment
which is a prime MIME candidate for the extended character set support.
I am managing the Danish part of EUnet, and I am very active on the
UNIX user group scene here, so I communicate a lot on email and news
etc in Danish, as well as other network services we are providing
here. And I am also quite active international standardization
in internationalization and character set support. I should thus be
in a good position to watch what is going on, with a bias to the
more advanced pro-MIME side. The danish language has 3 vowels ae o/ aa
outside the ASCII range of letters, and the use of these total
about 4 % of the letter use in the Danish language. Not doing
them right is considered incorrect and doing something like the
Quoted-Printable encoding does great damage to the ease of reading
text for non-MIME people.
Thus my observation is that although MIME works fine between any two
MIME users, it does not work that well between a MIME user and a
non-MIME user. MIME then tends to be used accordingly, only between
MIME users, and there is thus a tendency to create a two-camp
world, MIME and non-MIME.
Considering the cost statements brought forward earlier on this
list, that programmer time is more expensive than CPU time, I would
like to remark that users time, spent on problems with reading
text, is a cost much more important than both CPU and programmer time.
I am sure that MIME works, and a lot of other schemes would work also,
like Otha-san's, or some ISO 10646 encoding - if the sender and receiver
run compatible software, then there is no problem! The art of our
Internet design of new specifications is then to make it compatible
with the old.
That was why I created the mnemonic system some 4 years ago,
which has a smooth migration path from 7-bit ASCII thru 8-bit
character sets into 16-bit and even 32 bit encoding. It is described
in RFC1345. Our experience here with a customer base of 200 small
to large organisations, is that it works. We use it all the time
and we are not flamed by our customers. We have a united internet here!
My recommendation is thus that we, when doing enhanced character set
support for news, then use the mnemonics scheme as our basic exchange
plain text format. This is currently also possible in MIME, and we should
use it there too.
Keld