Re: French encoding [Was: Chapter on character sets]

From: Alain LaBonté  (
Date: Fri Jun 16 2000 - 16:25:28 EDT

À 12:37 2000-06-15 -0800, a écrit:
>The requirement for Euro was definitely very important, but as I remember
>the discussions, it was only with very great difficulty that any examples
>of Finnish text was produced. Much of the impetus for this was also a
>desire by some people to replace Latin-1 with what was audaciously called
>"Latin-zero". I'm thankful that that never happened - I can only guess at
>the mess that would have resulted.

[Alain] "Replace", no, as *we* wanted to respect history. A replacement
would have used the same number as before [Latin-1], as was done on the
different versions of MS-1252. I was among those people you're talining
about (and I am by chance currently in the room in Paris where this number
0 was first used, when I suggested it). "Latin-0" was just a "buzzword" to
hit imagination [in fact it made its effect perfectly], a code name. We
expected that it would not hold in ISO [which always counts starting in
"origin 1", not "0"]. It became "Latin-9" (curiously, in French it reads
"Latin neuf" ["neuf" means "9" and "brand new" at once in this context --
the words are exactly the same, as "latin" is masculin in French -- but in
this case the renumbering happened fortuitously], or ISO/IEC 8859-15 (I
sometimes tell myself it is 8859-1.5, the "1" that should have existed at
once [but the EURO SIGN did not exist in 1987, of course]!!!)

>it has been generated automatically for a while in data by WinWord (with
>the Canadian keyboard I type it dicrectly myself).

>Ah - now you've hit the nail on the head. What on earth was wrong with
>Unicode? For that is what WinWord uses to encode your favorite characters.

[Alain] Yes and no. Under the hood maybe. But it is not what is exchanged
-- the majority of our current applications on different platforms are
capable of exchanging 8-bit-fixed-length-precomposed data only (alas, we
have to live with this for a while). And that is the problem. Missing
French characters in ISO/IEC 8859-1 but which are [Bill, thank you for that
one!], in MS-1252 [and of course in Unicode], can't be exchanged with the
numerous applications we have on other platforms, in particular those using
EBCDIC CECP code pages. There was a missing link, it is undeniable, and we
filled the gap. We did our duty. It had to be done. It does not contradict
Unicode, on the contrary, it is a comforting element for the latter... At
least that is how the initial "Latin-0" team saw it... And we said it as

Alain LaBonté, project editor, ISO/IEC 8859-15 (Latin 9)
Paris (for a few hours)

This archive was generated by hypermail 2.1.2 : Tue Jul 10 2001 - 17:21:04 EDT