Re: Stroke encoding?

[Jacques Guy <jguy@xxxxxxxxxxxxxxxx>]

28/02/02 09:18:19, "Gabriel Landini" <G.Landini@xxxxxxxxxx> wrote:

>I see a problem with stroke encoding. Why? because 
>1. Kanjis are not always written top-bottom,

You mean that they are also written from left to right.
I don't see it as a problem. If you know Chinese you
can tell when the left-hand side is complete and the
next stroke belongs the right-hand side

>2. Some kanjis have radicals, 

_Most_, not some.

>which are in fact kanjis made out of more (smaller, 
>but not less complicated) kanjis.
>One would need spatial localisation of the strokes within 
>the "kanji space". Without this, the reader does 
>not know where the author was "writing".
>Just recording the strokes would be quite messy.

Recording them is straightforward. Reading them
back might be a bit messy. But not if you have
a letter that means "start from the top now"
or something similar.

Even in the case of "messy" characters like
wo3 ("I") I do not think that the sequence
of the strokes making it up can be mistaken
for another character. By the time you
have pie3, heng2, shu4-gou3, ti2, the first
three strokes, I think that you _know_
that it has to be wo3. I cannot think of
another character starting with that sequence.
But, admittedly, my Chinese is rusty.