NAME
   Lingua::JA::Sort::ReadableKey - Sorting and Romanizing Japanese

SYNOPSIS
     use Lingua::JA::Sort::ReadableKey;
     for ( map { $_->[0] }
          sort { $a->[1] cmp $b->[1] }
           map { [ $_, japanese_sortorder($_) ] } @utf8 ) {

DESCRIPTION
   First, does Lingua::JA::Sort::JIS do what you want? Look at that first.

   It may not do what you want if you want

   *  Kanji phrases sorted in their reading order, rather than as a
      separate block.

   *  A machine-readable or storable key so that comparisons and sorting
      can be done by a non-Japanese-aware system later.

   This module uses "Text::ChaSen" to do kanji-kana conversion, and then
   produces a comparable ASCII key for sorting.

   All text should be in "real" UTF-8 - that is, strings in Perl's internal
   format with the UTF-8 flag on.

 EXPORT
   The following methods are exported:

  kanji_to_kana
   Use ChaSen to convert a kanji sequence into hiragana. You obviously need
   to install ChaSen, and its Perl interface "Text::ChaSen" to make this
   work. You can get ChaSen from http://chasen.aist-nara.ac.jp/ and
   "Text::ChaSen" is bundled with it. If you have Debian, install the
   packages "chasen" and "libtext-chasen-perl". This code will work with
   both ChaSen1 and ChaSen2.

  japanese_pronunciation
   This turns a Japanese string into an ASCII representation of its
   reading. You can't sort on this, because Japanese don't sort according
   to the Latin alphabet, but you can use to label Japanese things for
   people who can't read Japanese. This will automatically call
   "kanji_to_kana" if necessary to get the reading of kanji strings.

  japanese_sort_order
   This returns an ASCII string which represents, in some bizarre magic
   encoding, the sort order of the Japanese input string, such that
   comparing the "japanese_sort_order" of two UTF-8 strings will tell you
   how they should be sorted in a Japanese dictionary.

   By "bizarre" and "magic", I mean that for each character, we find its
   order in the Japanese alphabet, and then replace that with
   "chr(32+$order)" so that it can be compared with "cmp".

SEE ALSO
   Lingua::JA::Sort::JIS, Text::ChaSen.

AUTHOR
   Simon Cozens, <[email protected]>

COPYRIGHT AND LICENSE
   Copyright (C) 2004 by Simon Cozens

   This library is free software; you can redistribute it and/or modify it
   under the same terms as Perl itself, either Perl version 5.8.4 or, at
   your option, any later version of Perl 5 you may have available.