From: Samuel Thibault (samuel.thibault@ens-lyon.org)
Date: Sat Jan 31 2009 - 17:43:14 CST
Jeroen Ruigrok van der Werven, le Sun 01 Feb 2009 00:16:50 +0100, a écrit :
> -On [20090131 23:32], Samuel Thibault (samuel.thibault@ens-lyon.org) wrote:
> >Here I do not really care about how things are pronounced, but what they
> >_mean_.
>
> But then still, a single kanji may have 1-5 meanings, and used in a compound
> it may get a whole different meaning.
Right, but that's the same for sighted people. The problem here is
that blind people don't know for sure _which_ kanji they are reading.
If they get to know that information, then they can infer the actual
meaning just like sighted readers do. I probably shouldn't have written
"meaning" in my previous mails, but "a description of which kanji it
is".
> So unless you use a morphological analyser (mecab, chasen), you are going to
> lose a lot of information if you insist on a raw one to one mapping from
> kanji to English.
Please re-read my initial mail: I'm precisely _not_ asking for a
one-to-one mapping to english (which CDU already provides actually),
but to something (in chinese, even) that tells which kanji it is, so
that the blind reader can then infer the meaning just like how sighted
readers do.
Samuel
This archive was generated by hypermail 2.1.5 : Sat Jan 31 2009 - 17:46:02 CST