Re: ISO/IEC 10646 versus Unicode

From: Alain LaBonté  (
Date: Fri Jul 19 2002 - 10:41:35 EDT

A 09:24 2002-07-19 -0400, James E. Agenbroad a écrit :
>JEA: To clarify, could some one say when the ISO 10646 effort began? Does
>it antedate Ed Hart's "ASCII and EBCDIC: Character set and code issues in
>systems application architecture" published by SHARE in June 1989? From
>its preface: "This activity started as a result of a challenge from the
>Human Factors Project at the SHARE 67 meeting in August, 1986."
>In ISO-IEC JTC1/SC2/WG2 N619R Date: 7 January 1991: "Document Register:
>N 1 Register of documents tabled at Kyoto and ad-hoc meeting
>N 2 Draft agenda of Kyoto ad-hoc meeting
>N 3 Initial membership and mailing list
>N 4R WG2 program of work (=97/2 N1436) and goals (=97/2 N1436 Annex)"
>I do not have these early WG2 documents which might have dates.

[Alain] When I joined JTC1/SC2 in 1987 (the year ISO TC 97 became ISO/IEC
JTC1), ISO/IEC 10646 was already under discussion.

    In 1991, I personally started a campaign -- in writing to all JTC1/SC2
national bodies on behalf of SHARE Europe (I was their manager of standards
at that time, after having been member of its executive board and their
only "American" member) -- to ask national bodies to try to reconsider the
mistake to have two standards instead of one, and to try to work with the
Unicode consortium to come to converging directions. My respected and
sincere friend, Mr. Dara Hekimi, was furious. I explained him that I
sincerely believed that, in spite of his very nice and consistent work in
ISO 2022 (and I have but admiration for this pioneer, who was a briliant
mathematicien apart from being the founder of ECMA), we ought to get less
complicated ways of doing things for such a huge chaarcter sets where
escaping would be extremely frequent, and not a mere exception, and would
lead to easy-to-anticipate problems of all kinds.

    The initial ISO/IEC 10646 was trying to be compatible in all its
256-characters pages, with the ISO/IEC 2022 code-extension-technique
structure, where 64 code positions were reserved for control characters.
SC2 was envisaging the implementation of the universal character set with
escape sequences in compatibility even with archaic 7-bit-only code
environments. This seemed to us, from SHARE Europe, a nonsense. We also
wished that the sacrificed bit of all octets be used, instead of
perpetuating the 7-bit-*octets* mascarade.

    Yes, ISO/IEC 10646 efforts started before the Unicode consortium was
created, as far as I know.

    That said, the Unicode approach was superior -- because simpler -- than
the first UCS drafts which went as far as DIS before being turned down.

Alain LaBonté

PS: SEAS (which became SHARE Europe in 1988) published a "white paper" on
the character coding mess in 1985. It published a new one (suggesting
solutions this time to IBM, including the "Davos requirement" of 1988 which
was the predecessor of what is now the ISO/IEC 1465 International Standard
[on character ordering], among other things) in 1990.

    Starting in 1988, if I remember well, SHARE Inc. (USA) and SHARE Europe
started close collaboration on this and the 1990 SHARE Europe white paper,
as well as the SHARE ÆCS (ASCII-EBCDIC c.s.) white papers have been
mutually influneced, thanks to Ed Hart who initiated the contacts. I'm very
glad he did, and we also all became best-of-friends.

This archive was generated by hypermail 2.1.2 : Fri Jul 19 2002 - 08:56:41 EDT