Ran Atkinson writes:
On Sep 18, 15:22, Keld J|rn Simonsen wrote:
} Subject: Re: 8-bit transmission in NNTP
Ran Atkinson writes:
% Well, it works for all of internet, in the raw mode. Only ASCII is
% assumed supported, and that can be assumed to be universally available,
% viz. ASCII being a requirement of RFC 822.
But it is NOT _mnemonic_ to HUMANS and hence of no real value !
Well, for Vietnamese it was designed to be mnemonic, and I think
it met its goal. All Vietnamese characters are supported, the letters
with one diacritic have two-char mnemonics, and the letters with
two diacritics have 3-char mnemonics. Some examples are:
A2 1ea2 LATIN CAPITAL LETTER A WITH HOOK ABOVE
a2 1ea3 LATIN SMALL LETTER A WITH HOOK ABOVE
A>' 1ea4 LATIN CAPITAL LETTER A WITH CIRCUMFLEX AND ACUTE
a>' 1ea5 LATIN SMALL LETTER A WITH CIRCUMFLEX AND ACUTE
A>! 1ea6 LATIN CAPITAL LETTER A WITH CIRCUMFLEX AND GRAVE
a>! 1ea7 LATIN SMALL LETTER A WITH CIRCUMFLEX AND GRAVE
A>2 1ea8 LATIN CAPITAL LETTER A WITH CIRCUMFLEX AND HOOK ABOVE
a>2 1ea9 LATIN SMALL LETTER A WITH CIRCUMFLEX AND HOOK ABOVE
A>? 1eaa LATIN CAPITAL LETTER A WITH CIRCUMFLEX AND TILDE
a>? 1eab LATIN SMALL LETTER A WITH CIRCUMFLEX AND TILDE
A(' 1eae LATIN CAPITAL LETTER A WITH BREVE AND ACUTE
a(' 1eaf LATIN SMALL LETTER A WITH BREVE AND ACUTE
A(! 1eb0 LATIN CAPITAL LETTER A WITH BREVE AND GRAVE
a(! 1eb1 LATIN SMALL LETTER A WITH BREVE AND GRAVE
A(2 1eb2 LATIN CAPITAL LETTER A WITH BREVE AND HOOK ABOVE
a(2 1eb3 LATIN SMALL LETTER A WITH BREVE AND HOOK ABOVE
A(? 1eb4 LATIN CAPITAL LETTER A WITH BREVE AND TILDE
a(? 1eb5 LATIN SMALL LETTER A WITH BREVE AND TILDE
The mnemonic here is: "a" and "A" stands for SMALL vs CAPITAL A,
"2" is HOOK, ">" is CIRCUMFLEX, "'" is ACUTE, "!" is GRAVE,
"?" is TILDE, "(" is BREVE. Turn 2>)? 90 degrees against the clock
and the mnemonic becomes quite intuitive.
(the second coloumn is the 10646 code.)
It may not be the preferred mnemonic of everyone, but it would not
be right to say that they are not mnemonic.
It certainly isn't better than QP as a general solution.
For Vietnamese I do not think there is a QP solution, so from a
standpoint that some support is better than no support, mnemonic is
far superior.
% Vietnamese is supported, as you know Ran.
Vietnamese encoding using Keld-char is UNREADABLE. Hardly "supported".
Well, this is a matter of taste. I have seen Vietnamese using RFC1345
mnemonics, so there are other opinions than yours on the case.
They are certainly more readable than any QP encoding of Vietnamese.
There is a convention for Vietnamese into ASCII which has been
used for many years on Vietnamese mailing lists and newsgroups, is
documented in an RFC, and is reasonably mnemonic for Vietnamese
readers/writers. Your encodings for Vietnamese are horrible and can't
be parsed by Vietnamese readers/writers. Your encodings are designed
to be tremendously Euro-centric and Danish-centric. They do not
represent a general solution to the problem and never have.
I admit that my encodings are not the best for Vietnamese, as you
should always be able to design something for a particular purpose,
eg Vietnamese, which is better for this particular purpose, than the
general solution. Mnemonic is a general solution, which covers all of
10646 (with Uxxxx codes for those chars without real mnemonics),
and can thus be used all over the world in the same way.
But the generality is then won on the expense of some beauty.
As discussed earlier, I don't think it is tremenduously Euro-centric
nor Danish-centric. Of cause I cannot remove my heritage when doing
brain-work, but a number of design rules have tried to remove this
culture bias. It was a design goal to be cultural independent.
This was attempted by trying to making the mnemonics look as the
originals, instead of using English (or Danish) words or abbreviations.
This is epecially true for the scientific characters, where symbols
like -> were preferred to "right-arrow". But also in the letter section
this was done, eg o/ was chosen instead of oe for the danish O WITH
STROKE, oe being the French LIGATURE OE. Also ISO transcription
standards were used when possible.
I admit that mnemonic is ASCII-centric, but that was because Internet
mail and other specs are ASCII-centric (viz. rfc822 and also the news
RFCs).
% It is quite widely implemented and deployed, I know many sites outside
% Denmark using it. It is not mainsteam, tho.
}-- End of excerpt from Keld J|rn Simonsen
It is not "widely implemented and deployed". It is more
deployed than it used to be but it is not widespread.
It is spead all over the world. It is part of IDA sendmail, which was,
and to some extend still is, "standard" sendmail. It is part of "pp"
which is the standard X.400 on internet, and also handles SMTP.
It is part of PMDF.
How widely it is spread I do not know, but making it into 3
prime SMTP MTAs and distributed for a couple of years certainly
makes it of more than just a Danish oddity.
Keld, on the subject of your rejected character set encoding
proposal you are without a doubt the most frequent and blatent liar
I've seen in over 10 years on the net. You truly astonish me.
I might write erroneous statements, but I do not do that with intention.
I am not the flaming kind, and I will let others judge the technical
merit of my contributions. People can also look in the smtp-ext
archives and see how gracefully Randall has handled me previously.
I still answer his mails, I sometimes wonder why.
Keld