----------------------------------------------------------------
  Making a new character pattern dictionary
    Written by H. Goto  (Rev. 20100212)
----------------------------------------------------------------

1. Preparation

  The following files are required.
  The <langcode> is the language code such as jpn.
  (Never include < > symbols.)

    cctable-<langcode> file.
    One or more TrueType font files for the language.

  You need to create your own cctable file.
  See cctable-format.txt for the details about the cctable file.


2. Making character image files

  Type as shown below to create character image files.

    $ makechar -font Font_1.ttf cctable-<langcode> Font_1.img
    $ makechar -font Font_2.ttf cctable-<langcode> Font_2.img
    ...

  Font_N.ttf is the name of TrueType file. Proportional fonts
  are required to support ligatures.
  Font_N.img is the name of the output image file.

  If you see a Warning message during the makechar process, you
  should avoid using the font file or remove the character which
  causes the warning.

  Each character image file contains a list of 128x128-pixel
  character images in binary form.


3. Making vector files

  Type as shown below to convert the character image files into
  vector files.

    $ makevec -F PLM Font_1.img Font_1.vec
    $ makevec -F PLM Font_2.img Font_2.vec
    ...

  Font_N.vec is the name of the output vector file.


4. Making a dictionary

  Type as shown below to create a dictionary.

    $ makedic -F PLM -o PLM-<langcode>.dic Font_*.vec


5. Copy the new dictionary

  Copy cctable-<langcode> and PLM-<langcode>.dic into share/
  directory under the NHocr installation path.


--
