Re: New Draft ISO 8859-0

From: Alain LaBont/e'/ SCT (alb@riq.qc.ca)
Date: Mon Jul 07 1997 - 16:33:50 EDT


Erland Sommarskog writes on comp-software-international@news-digests.mit.edu:

>Markus Kuhn <kuhn@cs.purdue.edu> skriver:
>>They want to replace the frankly quite useless characters:
>>
>>A4 CURRENCY SIGN
>>A8 DIAERESIS (standalone, spacing character)
>>B4 ACUTE ACCENT (standalone, spacing character)
>>B8 CEDILLA (standalone, spacing character)

[Erland] :
>They're also suggesting dropping the following six:
>
> 10/06 BROKEN BAR
> 11/02 PLUS OR MINUS SIGN
> 11/12 VULGAR FRACTION ONE QUARTER
> 11/13 VULGAR FRACTION ONE HALF
> 11/14 VULGAR FRACTION THREE QUARTERS
> 10/02 CENT SIGN
>
>Broken bar is a character I've never understood why it exists, and
>possibly the cent sign is little used. But the other four certainly
>looks reasonable. There's even ± key on my keyboard. On the other
>hand, who is using the macro[n] sign (¯) or or that matter the logical
>not (¬)?

[Alain] :
The macron sign should in fact have been called the overline character (as
in Latin 1 it is not, generally speaking, used as a diacritic [in fact, to
artificially represent the macron diacritical mark, I have myself, on
several occasions, used the *underline* character on the line above a
letter rather than the spacing macron of ISO/IEC 8859-1 for scholar French
-- to represent words borrowed from Arabic, for example
 _
qat ]). That said, the Latin 1 "macron" deserves other purposes: it is the
complement of the underline character.

It is used to make *primitive* box drawing, as in the following:
_______
EXAMPLE
¯¯¯¯¯¯¯
or
_______
¯¯¯¯¯¯¯
EXAMPLE
_______
¯¯¯¯¯¯¯

And the NOT SIGN is used in PL/1.

It also appears that the currency sign is used for certain control stuff in
some environments (as a field separator in certain packages, or, at British
Airways [if my memory is good], as a SUBTOTAL indicator). But that is not a
problem any more, we needed 2 characters less than the initial proposal (to
finally correct Latin 1 for 3 missing French characters and 4 missing
Finnish characters) and at this moment we sacrificed the PLUS-MINUS SIGN
for the EURO SIGN and kept the CURRENCY SIGN and the CENT SIGN intact.

Now for the PLUS-MINUS sign: it is not perfectly ideal that we give it up
in favour of the SPUTNIK-like CURRENCY SIGN, but it can use +- as a
fallback. And there is always a possibility to continue to use Latin 1 for
specific purposes which do not already use the EURO sign. For new
applications the fallback can be used. This appears reasonable.
Of course for reallocating a full table, something has to be sacrificed.
But impact should be minimal, that's what we did in all good faith and good
will, conscious that not absolutely everybody will be satisfied. However it
is a new part of ISO/IEC 8859, not modifying Latin 1 which continues to
exist, not for EURO SIGN usage, though.

Alain LaBonté
Iraklion

PS: I have all the sacrificed characters on my own keyboard, which is *the*
Canadian keyboard standard (CAN/CSA Z243.200-1992, based on ISO/IEC 9995-3
for international character usage), but they can easily be sacrificed in
practice for something more useful *in the 8-bit world*. In ISO/IEC 9995-3
all the sacrificed characters plus all the new ones, except EURO SIGN, can
be entered... this is excellent for ISO/IEC 10646... although this ISO
keyboard standard (as well as the Canadian standard, due for revision) will
have to be modified for EURO SIGN.

The new ISO/IEC 8859 part does not intend to compete with the UCS but
rather to rectify things in the 8-bit world and create links with mainframe
[and other environment] applications (which need to build conversion tables
from legitimate standard character sets -- for very good technical reasons,
in particular in the predominent EBCDIC market world which can't implement
in the 8-bit world character sets with more than 191 graphic characters --
à la Windows [in itself a goodie, but with a missing link]). This is *the*
missing link, so that the EURO and two official European languages can be
used integrally in a standard way (: (;



This archive was generated by hypermail 2.1.2 : Tue Jul 10 2001 - 17:20:35 EDT