Jungshik Shin wrote: > > It's unnecessary to handle ALL cases. You could address only issues > > encountered/expected by your end users. IMHO, it is more important > > to make an application be light-weight and run in 99% cases. Or, you > > may find your language used by, say, 10000 people, and none uses the > > extra features that you spend 40% of your development labour. > > As you wrote, one can do what one believes. Anyway, correctly > handling non-BMP characters are not so much difficult (40% of your > devel. time for 1% constituency seems to me too big an exaggeration > :-) I know you're just maing your case clear...). Moreover, with > Math characters in plane 1 and MathML more widely used, it'd not be > so rare to find people who want to use non-BMP characters.
Just checked the Unicode NamesList.... You really hit me on the point. Then it seems 16 bits is really not enough in a future not far away. Though, I still do not like the idea to have 4 bytes to store an ASCII character. But ... I must reconsider my point. Best regards, Wu Yongwei -- Linux-UTF8: i18n of Linux on all levels Archive: http://mail.nl.linux.org/linux-utf8/
