On Mon, Nov 15, 2004 at 07:33:13PM +0800, Xie Qian wrote: > > > * fontconfig-2.2.3-alllang-20041114.patch > > Some asian fonts include fonts more than one language > > family, i.e. ttc fonts > > Considering the Uni-Han problem, no CJK fonts can be > acceptable in China mainland, Taiwan, Japan and Korea due to > character shape diversity. Exclusive lang tag can be use to > match most suitable CJK font. So this is a designed feature. Personally, I find the "shape difference" argument (in the Han unification process of Unicode) very weak. Some months ago, I thought of modifying some public-licensed CJK font. In the process, I found that Unicode assigned different code points for Chinese and Japanese for some characters, but if a radical is added, there is only one code point. This makes it impossible to design fonts for more than one language. (This certainly should be possible, since we Chinese are not so picky about the shape differences, but the Japanese are more picky, so we can just pick the so-called "Japanese" shapes when they differ "too much".) Also in the process, I found a number of different glyphs (obviously for "shape differences" between different languages) where I cannot notice any difference at all even if I look very hard, or the difference is so small I can't believe even the Japanese will complain. My own conclusion is, to the Chinese, the "shape difference" is an artificial myth. (After all, all these different shapes are valid in Chinese, either as an old form, a variant form, or as a "popular but written incorrectly" form.) Personally, I think the "shape difference" criteria in the Han unification process is misguided and counter-productive in practice.
