As I have said for some time now and backed myself up in fairly
clear technical terms, the document "draft-ietf-822ext-mnemonics.txt"
has significant technical flaws that make its use outside of European
alphabetic languages inappropriate in its present form.
well, we had some conversation on this some time ago, but
I did not get any response to my comments then.
Maybe some mail was lost? If you have time we could discuss it
offline, if you feel it mostly like a repetition.
Even alphabetic Asian languages, such as Vietnamese, are not well
served or equitably treated by this document and its author Keld
Simonsen has explicitly rejected my specific proposals to change it so
that it would treat at least all alphabetic languages on equal terms.
Vietnamese is a language well known for its many accents. I might
be able to make a scheme which can accomodate vietnamese as well as
other languages based on the latin script. I have tried before
to do Vietnamese support, but evidently not to the satisfaction
of those people in need for it.
I do not remember any specific proposals from you on Vietnamese, Ran.
Please forward them to me if you have some. The last part of
our conversation that I remember was that I doubted if the current
Vietnamese trend of very specific post-accents and usage of
variant characters could be merged, thus challenging you to
come forward with such a proposal.
I think that it would be fine as an experimental protocol or as an
informational document, possibly with a registered type with the IANA.
I strongly oppose any and all attempts to put it on the standards
track until the author makes the changes that I and others have
outlined that are necessary to treat all alphabetic languages on an
even basis and to permit multiple character set profiles to work using
the same mechanism.
One design goal has actually been to treat every *character set*
and not language, equally. That is, all 7- or 8-bit character sets
in the ECMA registry can actually be written with two-char mnemonics.
There are of cause some languages that get preferential treatment,
such as American, English, Dutch, Swahili and Hawaian, but this
has historic origins. Actually, if you use a character set which
supports your native language, you will get support for that
with the specifications, be it either Japanese, Danish or Vietnamese.
You only need mnemonic support for characters that are not in
the character set you use.
As said above I miss specific proposals on how to make it better.
I think I have missed completely your points on the multiple character
set profiles. I would like to hear more about that.
I think that the document "draft-ietf-822ext-charsets-*.txt" would
be very useful as an informational document and ery much hope that it
will be published informationally. It does not contain standards
track material and is not suitable for the standards track, and I
oppose any effort that might be made to put it on the standards track.
Thanks for your support!
Keld