Re: Unicode & Han

From: Martin J Duerst (
Date: Sun Aug 11 1996 - 12:58:19 EDT

Alain LaBonte' wrote:

>There has been a proposition by the Chinese delegates presented twice in
>SC2/WG2, if I remember well, that I found genial... To complement the
>existing set, they were proposing a composition-on-the-fly method that would
>have allowed to compose any character based on Chinese radicals.
This is actually worked on currently in the IRG. It looks genial, but it also
has some problems that one has to be very careful about.

>As a
>primitive and occasional user of Chinese, I would have been delighted by
>this: I do not have an intelligent input method for entering Cinese and
>sometimes I look hard to find some Han characters in the UCS. With this
>method I would not search more than a few seconds to retrieve the right
As you should know, coding methods and input methods are not related
one-to-one. You can easily have the input method you immagine with
current codepoints.

>In fact it would also simplify sorting and ordering: right now CJK
>characters are inherently sorted; however as we can't insert extra
>characters in the table, they will be added in other planes and we'll have
>to have complex methods for sorting them (or huge tables). The composition
>method would not require such complex processing at all for
>radical/strokecount order.

Radical/strokecount is not the only sorting order. Usually, phonetic order
is preferred, but that needs additional information. Also, it is not in general
possible to derive the (traditional) radical from the decomposition.
The best example, which easily surprises most Chinese and Japanese,
is U+548C.

Regards, Martin.

This archive was generated by hypermail 2.1.2 : Tue Jul 10 2001 - 17:20:31 EDT