Hi, At Tue, 02 Apr 2002 18:08:10 +0100, Markus Kuhn wrote:
> Again, that's just the old many-to-one issue here, nothing critical. > The fact that Unicode contains both > > U+00B5 MICRO SIGN > U+03BC GREEK SMALL LETTER MU > > (both of which are really the exact same character in most people's > view) didn't prevent ISO 8859-1 being mapped to UCS by asigning its 0xB5 > to U+00B5 MICRO SIGN in the round-trip compatibility table. Now I am not saying about this problem. > The compatibility characters are there in Unicode to allow you to chose > to either use the unification rules of JIS or the unification rules of > the IRG, at your choice. Well, when I use Unicode, I have to obey Unicode's unification rule. When I use JIS, I have to obey JIS's unification rule. It is *always*. COMPATIBILITY IDEOGRAPH exists only for round-trip conversion, not for different unification rule. For example, IBM's table says that: SJIS UCS 8A43 6D77 (in CJK UNIFIED IDEOGRAPHS region) EBA2 FA45 (in CJK COMPATIBILITY IDEOGRAPHS region) This is an example of a pair of characters in SJIS's viewpoint and a pair of glyphs of one character in Unicode's viewpoint. See page 843 of http://unicode.org/charts/PDF/UF900.pdf . Usage of SJIS 8A43 means that the character is *not* SJIS EBA2. However, UCS 6D77 is a unification of SJIS 8A43 and SJIS EBA2, and thus, usage of UCS 6D77 cannot specify SJIS 8A43's glyph. It is totally legal to use SJIS EBA2's glyph for UCS 6D77, because SJIS 8A43's glyph and SJIS EBA2's glyph are unified in Unicode. Usage of UCS FA45 can specify SJIS EBA2's glyph (and can exclude SJIS 8A43's glyph). However usage of COMPATIBLITY IDEOGRAPHS for other than compatibility purpose is discouraged. Thus, we should not use UCS FA45 to specify SJIS EBA2's glyph. Accidentally, Unicode Consortium's sample glyph for UCS 6D77 is same as SJIS EBA2 (and different with SJIS 8A43). Thus, glyphs of UCS FA45 and UCS 6D77 in the page 843 of http://unicode.org/charts/PDF/UF900.pdf are same, which may have confused you. --- Tomohiro KUBOTA <[EMAIL PROTECTED]> http://www.debian.or.jp/~kubota/ "Introduction to I18N" http://www.debian.org/doc/manuals/intro-i18n/ -- Linux-UTF8: i18n of Linux on all levels Archive: http://mail.nl.linux.org/linux-utf8/
