Joël Lopes Da Silva on the Web!

Since I discovered LaTeX, I have become a real fan of this typesetting system. Yet, something bothered me very much: I could not use easily Japanese with LaTeX using a normal encoding, that is to say using UTF-8. Using CJK is not too difficult, but getting CJKutf8 to work is much harder. I just couldn't resign myself to use some strange encoding popular in Japan, mostly for Windows Addicts...

But I just found how to do it. Since it is not so easy, I want to do a copy-paste how-to explaining precisely how to install UTF-8 Japanese support in MacTeX. So, just follow the guide...

Updated

• Friday, 14 November 2008: updated for MacTeX 2008 support on Leopard; small "bugfix" of the tutorial.

Of course, installing MacTeX is a prerequisite. This how-to has been tested on Mac OS X 10.5 (Leopard) with both versions 2007 and 2008 of MacTeX, and it is just the adaptation of a procedure described in a Debian Readme file (/usr/share/doc/latex-cjk-common/README.Debian.gz in Debian Etch)...

The problem when we want to type Japanese directly in UTF-8 encoding in a LaTeX document is that we have to use the package CJKutf8 instead of CJK, and by default, there is no font with Japanese symbols adapted for UTF-8 typesetting in LaTeX. Therefore, the main goal of this how-to is to install the font Cyberbit into your LaTeX distribution. Please, be sure to acknowledge the font's license.

First, in a terminal, download and unzip the font file:

% wget ftp://ftp.netscape.com/pub/communicator/extras/fonts/windows/Cyberbit.ZIP
% unzip Cyberbit.ZIP
% rm -f Cyberbit.ZIP

Then, we need to become root in order to prepare some directories and files:

% sudo -s
# mkdir -p /usr/local/share/fonts/truetype/bitstream/
# mv Cyberbit.ttf /usr/local/share/fonts/truetype/bitstream/cyberbit.ttf
# mkdir -p /usr/src/cyberbit-fonts/
# exit
% sudo chown "$USER":staff /usr/src/cyberbit-fonts/
% cd /usr/src/cyberbit-fonts/

/usr/src/cyberbit-fonts/ will be our build directory. We now must prepare some more files in this directory:

% ln -s /usr/local/share/fonts/truetype/bitstream/cyberbit.ttf
% cp /Library/TeX/Root/texmf-dist/source/latex/CJK/utils/subfonts/subfonts.pe .

% fontforge -script subfonts.pe cyberbit.ttf cyberbit \
      /Library/TeX/Root/texmf/fonts/sfd/Unicode.sfd

This will take a very long time, probably several hours, so if you have something else to do, go for it! Then you must generate a file called cyberbit.map:

% touch cyberbit.map
% for i in *.pfb
  do
    echo "$(basename $i .pfb) $(basename $i .pfb) <$i" >> cyberbit.map
  done

Then, we need root's priviledges for almost all the things left to do, so type the following command:

% sudo -s

in order to become root, and type your password. Then do the following:

# mkdir -p /usr/local/texlive/texmf-local/fonts/map/dvips/cyberbit/
# mkdir -p /usr/local/texlive/texmf-local/fonts/{afm,type1,tfm}/cyberbit/
# cp cyberbit.map /usr/local/texlive/texmf-local/fonts/map/dvips/cyberbit/
# cp *.afm /usr/local/texlive/texmf-local/fonts/afm/cyberbit/
# cp *.pfb /usr/local/texlive/texmf-local/fonts/type1/cyberbit/
# cp *.tfm /usr/local/texlive/texmf-local/fonts/tfm/cyberbit/
# echo "Map cyberbit.map" >> /Library/TeX/Root/texmf-var/web2c/updmap.cfg
# cd ..
# export PATH="/usr/texbin:$PATH"
# texhash
# updmap
# updmap-sys

Then you have to replace a file of the distribution MacTeX by another one:

# rm -f /Library/TeX/Root/texmf-dist/tex/latex/CJK/UTF8/c70song.fd
# mkdir -p /usr/local/texlive/texmf-local/tex/latex/CJK/UTF-8/
# cd /usr/local/texlive/texmf-local/tex/latex/CJK/UTF-8/
# touch c70song.fd && chmod -R go+w .
# open -a TextEdit c70song.fd

Now, a new TextEdit window should have opened, with an empty file. Copy-paste the following text into the empty file:

Then save the file. Lastly, run:

# chown -R root:wheel . && chmod -R go-w .
# texhash

That's it! You should now be able to compile with pdflatex such a file (encoded in UTF-8, check out in your LaTeX editor's preferences):

\documentclass[a4paper,12pt]{article}

%%%%%%%%%%%%%%%%%%%%%%%
% Encodings, fonts... %
%%%%%%%%%%%%%%%%%%%%%%%

\usepackage[utf8]{inputenc}
\usepackage[T1]{fontenc}


%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
% Oriental languages specificic options %
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

\usepackage[T1]{CJKutf8}

\usepackage[overlap,CJK]{ruby}	  % Furigana support
\renewcommand{\rubysize}{0.5}	  % Furigana size
\renewcommand{\rubysep}{-0.3ex}	  % Spacing between Furigana and Kanji


%%%%%%%%%%%%%%%%%
% Title, author %
%%%%%%%%%%%%%%%%%

\title{Japanese in UTF-8 with \LaTeX}
\author{Joel \textsc{Lopes Da Silva}}


\begin{document}

	\begin{CJK}{UTF8}{song}

		\maketitle
		\thispagestyle{empty}

		\section{タイトル}

			おめでとうございます。\ruby{自分}{じぶん}でできましたね。

			\bigskip

			これから、日本語で \LaTeX の\ruby{資料}{しりょう}を\ruby{書}{か}けます。

		\section{Another title}

			You can even use English words within a 日本語の\ruby{資料}{しりょう}
			~(document in Japanese).

			\bigskip

			Have fun!

	\end{CJK}

\end{document}

Compiling this basic LaTeX document with pdflatex should result in such a file. Enjoy!