On Sun, 25 Nov 2001, Philipp Reichmuth wrote: AA> On the contrary i have seen thousands os applications and web AA> pages using Unicode for Chinese and Japanese inspite of these AA> language scripts requiring large use of the Unicode encoding AA> space.
As a Korean whose script takes up an unnecessarily huge chunk of code space with precomposed syllables in BMP and has many features in common with Brahmi-derived scripts, would you reconsider your 'conspiracy theory' if I tell you that I strongly believe that ISCII and Unicode have done a much better job with Indic scripts than Korean standard body did with Korean script? Assigning precomposed characters or presentation forms (like half-form) their own code points may have helped some more web pages than otherwise be put up using technology available in mid-90's. However, as others including Philipp have told you many times, it's like buying a short-term advantage of being able to get away with a quick and dirty trick ( proprieatary fonts and incompatible encoding varieties ) at the expense of being flexible, extensible, amenable to natural language processing, suitable for DB application and many other advantages. PR> In theory, it would have been possible to do this for Chinese or PR> especially Korean as well, but there it would be even more complicated PR> to implement a font, and data would have become a lot larger as Not only in theory but also in practice, Koeran has been moving in that direction using Conjoining Jamos at U+1100 because that's the *only* viable way to represent Korean script without any artificial restriction imposed by state-of-the art of 1990's. Perhaps a few decades from now people may consider 1990's as a peculiar period in terms of Korean script representation in computer. (In pre-KS C 5601-1987 standard, 'jamos' were used to represent syllables with SI and SO to toggle between US-ASCII and Korean) PR> to implement a font, and data would have become a lot larger as Designing good looking fonts for Korean certainly takes a lot of effort, but that difficulty has become less to do with the way Korean script is represented in storage with the advent and wide deployment of OT and other smart font/rendering technologies. Data size is for sure a concern, but I think it'll be less of an issue compared with significant gains of going that way as time goes by (party due to things like GMR and CMR :-) among other things ) PR> compared to unicode if one had to encode each character by radicals Compared to Unicode?? Encode each character by radicals?? For Chinese characters, you're right because radical-based-encoding is not a part of Unicode/10646 (yet) and might never be, but for Korean, Unicode/10646 does have provisions for representing Korean scripts using *consonants and vowels* instead of syllables. That's what I'm talking about in the above. Jungshik Shin
