home *** CD-ROM | disk | FTP | other *** search
- Path: sparky!uunet!cs.utexas.edu!sun-barr!sh.wide!wnoc-tyo-news!cs.titech!titccy.cc.titech!necom830!mohta
- From: mohta@necom830.cc.titech.ac.jp (Masataka Ohta)
- Newsgroups: comp.std.internat
- Subject: Re: The ultimate encoding
- Message-ID: <2769@titccy.cc.titech.ac.jp>
- Date: 21 Jan 93 17:02:58 GMT
- References: <1iddeeINN58g@rodan.UU.NET> <1993Jan9.220818.25882@enea.se> <2677@titccy.cc.titech.ac.jp> <1993Jan10.103419.2364@enea.se>
- Sender: news@titccy.cc.titech.ac.jp
- Organization: Tokyo Institute of Technology
- Lines: 36
-
- In article <1993Jan10.103419.2364@enea.se>
- sommar@enea.se (Erland Sommarskog) writes:
-
- >>If an application must neglect language difference, it can do so in the
- >>application.
- >
- >It must equalize all the umpteen instances, yes. And it better do if
- >the user is not going to be violent when he is told "you got the wrong
- >sort of 'W' here, please replace".
-
- What? How do you think about a system which distigushes 'A's in
- Latin/Greek/Cyrillic and tells "you got the wrong sort of 'A' here,
- please replace"?
-
- The system is just poorly desinged.
-
- >But of course the application still
- >has to follow the language rules preferred by the user.
-
- Applications just says there is something wrong in applications' own way.
- In good applications, the error messages should be smart. Then, users
- correct errors as they like. That's all.
-
- >But then what's
- >the point with all these umpteen instances?
-
- For almost all purposes, there is only one correct instance.
-
- But, sometimes, it is necessary to make ambiguious serach of, say,
- personal names in Han characters. Then, the language information should
- be dropped. BUT, at the same time, it is necessary to unify many characters
- which have different code points in Unicode. Many Han characters which
- have quite different forms share the common origin and should be
- considered equal for the ambiguious search.
-
- Masataka Ohta
-