rja@edison.GE.COM (rja) (03/15/88)
Both AT&T and HP provide systems which support full Asian languages such as Japanese using both 8-bit (for Kana as a superset of ASCII) and 16-bit chracters (for Kanji, which are themselves a subset of Chinese Han Zi). The AT&T system meets all of the relevant standards for both Japanese and Chinese. I had thought that the HP system was incompatible, but a recent posting from Dave Decot in comp.unix.wizards seems to indicate that HP is also following the Japanese JIS standards. (I tried to send e-mail to him, but the HP gateway couldn't find him and bounced on me). The AT&T Sys V.3.1 KSH-I (Internationalised Korn Shell) was modified to remove 7-bit character dependencies. HP has done similar work. Hence, it is possible to remove these implicit dependencies in UNIX & C. There is a need to be able to handle these 16-bit and 8-bit characters in C intelligently. I've yet to be convinced how best to do it. This SHOULD be looked into by the ANSI committee until a consensus can be reached on how to do this. I don't believe that the std itself should be held up just because of longchar support. These views are my own and should not be construed as being those of my employer GE-Fanuc.
gwyn@brl-smoke.ARPA (Doug Gwyn ) (03/20/88)
In article <1386@edison.GE.COM> rja@edison.GE.COM (rja) writes: > There is a need to be able to handle these 16-bit and 8-bit characters >in C intelligently. I've yet to be convinced how best to do it. This >SHOULD be looked into by the ANSI committee until a consensus can be reached >on how to do this. I don't believe that the std itself should be held up >just because of longchar support. ??? This HAS been looked into by X3J11 and a solution proposed, that has been accepted by a large segment of the internationalization folks. I happen to dislike the solution because it treats character sets such as kanji as second-class citizens, but the Japanese seem to not mind. My guess is that they're so desperate for a solution that they'll accept anything workable.