L2/02-054

Title: Error in canonical mapping for U+F951
From: Ken Whistler

An error has been reported in the canonical mapping for
U+F951, one of the KS C 5601-1997 compatibility ideographs.
The error is blatantly obvious, if you examine the entry
on p. 798 of TUS 3.0.

The source of the error is a typo ("96FB" instead of the
correct "964B") entered into UnicodeData.txt.

The problem, however, is that this typo, however obvious
in retrospect, involves a canonical mapping in the
standard. And it is a canonical mapping not for a newly
added compatibility character in Unicode 3.2, but rather
for one that has been around since the hazy old days of
Unicode 1.0. The canonical mapping itself was added as
of Unicode 2.1.5, so it straddles the critical Unicode 3.0
and 3.1 boundaries for normalization stability. And
the wrong mapping is now enshrined in the normative
test for normalization:

birdie:kenw/work/unicode/staging320> egrep F951 NormalizationTest-3.2.0d6.txt
F951;96FB;96FB;96FB;96FB; # (ï¥; é»; é»; é»; é»; ) CJK COMPATIBILITY IDEOGRAPH-F951

In my opinion, this obvious error should be corrected,
but since it would involve another exception to the
guarantee of normalization stability, it clearly
would need a formal decision by the UTC (and have to
be treated with equal seriousness as the normalization
corrigendum for U+FB1D in Unicode 3.1).

What sayeth the committee?

==========================================================

----- Begin Included Message -----

> -----Original Message-----
> From: SADAHIRO Tomoyuki [mailto:bqw10602@nifty.com]
> Sent: Saturday, February 02, 2002 7:26 PM
> To: errata@unicode.org
> Subject: Beta Bug Report: an erroraneous mapping compat. ideograph
> 
> Hello, Unicode masters.
> 
> UnicodeData-3.2.0d8.txt says
> 
> F951;CJK COMPATIBILITY IDEOGRAPH-F951;Lo;0;L;96FB;;;;N;;;;;
> 
> but it should be
> 
> F951;CJK COMPATIBILITY IDEOGRAPH-F951;Lo;0;L;964B;;;;N;;;;;
> 
> http://www.unicode.org/cgi-bin/GetUnihanData.pl?codepoint=F951
> http://www.unicode.org/cgi-bin/GetUnihanData.pl?codepoint=964B
> http://www.unicode.org/cgi-bin/GetUnihanData.pl?codepoint=96FB
> 
> Regards,
> SADAHIRO Tomoyuki
> E-mail: bqw10602@nifty.com

----- End Included Message -----