| リビジョン | d2fc66f97142cda1a1c89baa11e18df6b5e0bc1b (tree) |
|---|---|
| 日時 | 2011-11-22 07:35:00 |
| 作者 | KUROKI Yusuke <kuroky@user...> |
| コミッター | KUROKI Yusuke |
Updated the draft for post-proceedings.
doc/ajt-devel-ltja.tex (diff)| @@ -1,1230 +1,1238 @@ | ||
| 1 | -%#!lualatex ajt-devel-ltja | |
| 2 | -\documentclass{ajt} | |
| 3 | - | |
| 4 | -%%% Packages used in this paper | |
| 5 | - | |
| 6 | -%%% Font setting for \LuaTeX; this is extract from ajt.cls | |
| 7 | -\makeatletter | |
| 8 | - \if@print | |
| 9 | - \RequirePackage{fontspec,xunicode} | |
| 10 | - \RequirePackage{luatextra} | |
| 11 | - \setmainfont[Mapping=tex-text]{Palatino LT Std} | |
| 12 | - \setsansfont[Mapping=tex-text]{Optima LT Std} | |
| 13 | - \else | |
| 14 | - \RequirePackage{fontspec,luatextra} | |
| 15 | - \setmainfont[Mapping=tex-text]{TeX Gyre Pagella} % \simeq Palatino | |
| 16 | - \fi | |
| 17 | - | |
| 18 | -%%% LuaTeX-ja | |
| 19 | -\usepackage{luatexja,luatexja-fontspec} | |
| 20 | -\ltjsetparameter{jacharrange={-3,-8}} | |
| 21 | -\DeclareFontShape{JY3}{mc}{m}{n}{<-> s*[0.92489] file:ipam.ttf:jfm=ujis}{} | |
| 22 | -\DeclareFontShape{JY3}{gt}{m}{n}{<-> s*[0.92489] file:ipag.ttf:jfm=ujis}{} | |
| 23 | -% quick hack: monospaced Japanese font by \ttfamily | |
| 24 | -\DeclareKanjiFamily{JY3}{\ttdefault}{}{} | |
| 25 | -\DeclareFontShape{JY3}{\ttdefault}{m}{n}{<-> s*[0.92489] file:ipag.ttf:jfm=mono}{} | |
| 26 | - | |
| 27 | - | |
| 28 | -%%% LTXexample environment | |
| 29 | -\usepackage{showexpl,lltjlisting} | |
| 30 | -\lstset{basicstyle=\ttfamily\small, width=0.3\textwidth, basewidth=.5em} | |
| 31 | - | |
| 32 | -%%% Verbatim environment | |
| 33 | -\usepackage{fancyvrb} | |
| 34 | -\CustomVerbatimEnvironment{code}{Verbatim}% | |
| 35 | -{numbers=left,xleftmargin=1.5em,baselinestretch=1.069,fontsize=\small} | |
| 36 | -\CustomVerbatimEnvironment{codewithoutnum}{Verbatim}% | |
| 37 | -{xleftmargin=1.5em,baselinestretch=1.069,fontsize=\small} | |
| 38 | -\CustomVerbatimEnvironment{codewithoutnumsmall}{Verbatim}% | |
| 39 | -{xleftmargin=1.5em,baselinestretch=1.0,fontsize=\footnotesize} | |
| 40 | -\DefineShortVerb{\|} | |
| 41 | - | |
| 42 | -%%% Others | |
| 43 | -\usepackage{mflogo,booktabs} | |
| 44 | -\definecolor{grayx}{gray}{0.85} | |
| 45 | - | |
| 46 | -%%% Mandatory article metadata %%% | |
| 47 | -\title{Development of \LuaTeX-ja package} | |
| 48 | -\author{Hironori Kitagawa {\normalsize 北川 弘典}} | |
| 49 | -\address{\LuaTeX-ja project team} | |
| 50 | -\email{h\_kitagawa2001@yahoo.co.jp} | |
| 51 | - | |
| 52 | -\keywords{\TeX, p\TeX, \LuaTeX, \LuaTeX-ja, Japanese} | |
| 53 | -\abstract{% | |
| 54 | -\LuaTeX-ja package is a macro package for typesetting Japanese | |
| 55 | -documents under \LuaTeX. The package has more flexibility of | |
| 56 | -typesetting than \pTeX, which is widely used Japanese extension of \TeX, | |
| 57 | -and has corrected some unwanted features of \pTeX. | |
| 58 | -In this paper, we describe specifications, the current status and some | |
| 59 | -internal processing methods of \LuaTeX-ja. | |
| 60 | -} | |
| 61 | - | |
| 62 | -\newcommand{\parname}[1]{\textsf{#1}} | |
| 63 | -\newcommand{\jstrut}{\vrule width0pt height\cht depth\cdp} | |
| 64 | -\newcommand{\imagfm}[1]{\ifvmode\leavevmode\fi% | |
| 65 | - \hbox{\fboxsep=0pt\fbox{\setbox0=\hbox{#1}\copy0\kern-\wd0 | |
| 66 | - \smash{\vrule width \wd0 height 0.4pt depth0.4pt}}}} | |
| 67 | -\begin{document} | |
| 68 | - | |
| 69 | -%%% Do not forget to start with \maketitle! | |
| 70 | -\maketitle | |
| 71 | - | |
| 72 | -\section{Introduction} | |
| 73 | -\subsection{History} | |
| 74 | -To typeset Japanese documents with \TeX, ASCII \pTeX~\cite{ptex} has | |
| 75 | -been widely used in Japan. There are other methods---for example, using | |
| 76 | -Omega and OTP~\cite{omega}, or with the CJK package---to do so, however, | |
| 77 | -these alternative methods did not become majority. The author thinks | |
| 78 | -that this is because \pTeX\ enables us to produce high-quality documents | |
| 79 | -(e.g.,~supporting vertical typesetting), and the appearance of \pTeX\ is | |
| 80 | -earlier than that of alternatives described above. | |
| 81 | - | |
| 82 | -However, \pTeX\ has been left behind from the extensions of \TeX\ such | |
| 83 | -as \eTeX\ and \pdfTeX, and the diffusion of UTF-8 encoding. In recent | |
| 84 | -years, the situation has become better, by development of | |
| 85 | -|ptexenc|~\cite{ptexenc} by Nobuyuki Tsuchimura (\hbox{土村展之}), | |
| 86 | -$\varepsilon$-\pTeX~\cite{eptex} by the author,~and u\pTeX~\cite{uptex} | |
| 87 | -by Takuji Tanaka (田中琢爾). However, continuing this approach, namely, | |
| 88 | -to develop an engine extension localized for Japanese, is not wise. This | |
| 89 | -approach needs lots of work for \emph{each} engine. In addition, if we | |
| 90 | -use \LuaTeX, the necessity of an engine extension is getting smaller | |
| 91 | -because \LuaTeX\ has an ability to hook \TeX's internal process by using | |
| 92 | -Lua callbacks. | |
| 93 | - | |
| 94 | - | |
| 95 | -There were several experimental attempts to typeset | |
| 96 | -Japanese documents with \LuaTeX\ before. Here we cite three examples: | |
| 97 | -\begin{itemize} | |
| 98 | -\item |luaums.sty|~\cite{luaums} developed by the author. This | |
| 99 | - experimental package is for creating a certain Japanese-based presentation | |
| 100 | - with \LuaTeX. | |
| 101 | -\item the \emph{luajalayout} package~\cite{luajalayout}, formerly known as the | |
| 102 | - \emph{jafontspec} package, by Kazuki Maeda (前田一貴). This package is based on | |
| 103 | - \LaTeXe\ and \emph{fontspec} package. | |
| 104 | -\item the \emph{luajp-test} package~\cite{luajp-test}, a test package made by | |
| 105 | - Atsuhito Kohda (香田温人), based on articles on the web page~\cite{joylua}. | |
| 106 | -\end{itemize} | |
| 107 | -However, these packages are based on \LaTeXe, and do not have much | |
| 108 | -ability to control the typesetting rule. And it is inefficient that more | |
| 109 | -than one people separately develop similar packages. Development of the | |
| 110 | -\LuaTeX-ja package is started initially by the author and Kazuki Maeda, because of | |
| 111 | -these situations. | |
| 112 | - | |
| 113 | -\subsection{Development policy of \LuaTeX-ja} | |
| 114 | -\label{ssec-pol} | |
| 115 | -The first aim of \LuaTeX-ja project was to implement features (from the | |
| 116 | -`primitive' level) of \pTeX\ as macros under \LuaTeX, therefore \LuaTeX-ja is | |
| 117 | -much affected by \pTeX. However, as development proceeded, some | |
| 118 | -technical/conceptual difficulties arose. Hence we changed the aim | |
| 119 | -of the project as follows: | |
| 120 | -\begin{itemize} | |
| 121 | -\item\emph{\LuaTeX-ja offers at least the same flexibility of | |
| 122 | - typesetting that p\TeX\ has.} | |
| 123 | - | |
| 124 | - We are not satisfied with the ability of producing outputs conformed to | |
| 125 | - JIS~X~4051~\cite{jisx4051}, the Japanese Industrial Standard for | |
| 126 | - typesetting, or to a technical note~\cite{w3c} by W3C; | |
| 127 | - if one wants to produce very incoherent outputs for some reason, it | |
| 128 | - should be possible. | |
| 129 | -In this point, previous attempts of Japanese typesetting with \LuaTeX\ | |
| 130 | - which we cited in the previous subsection are inadequate. | |
| 131 | - | |
| 132 | -\pTeX\ has some flexibility of typesetting, by changing internal | |
| 133 | - parameters such as |\kanjiskip| or |\prebreakpenalty|, and by using | |
| 134 | - custom JFM (Japanese TFM). Therefore we decided to include these | |
| 135 | - functionality to \LuaTeX-ja. | |
| 136 | - | |
| 137 | -\item\emph{\LuaTeX-ja isn't mere re-implementation or porting of \pTeX; | |
| 138 | - some (technically and/or conceptually) inconvenient features of | |
| 139 | - \pTeX\ are modified.} | |
| 140 | - | |
| 141 | - We describe this point in more detail at the next section. | |
| 142 | -\end{itemize} | |
| 143 | - | |
| 144 | - | |
| 145 | -\subsection{Overview of the processes} | |
| 146 | -\label{ssec-over} | |
| 147 | -We describe an outline of \LuaTeX-ja's process in order. | |
| 148 | - | |
| 149 | -\begin{itemize} | |
| 150 | -\item In the |process_input_buffer| callback: treatment of breaking | |
| 151 | - lines after a Japanese character (in Subsection~\ref{ssec-line}). | |
| 152 | - | |
| 153 | -\item In the |hyphenate| callback: font replacement. | |
| 154 | - | |
| 155 | -\LuaTeX-ja looks into for each \textit{glyph\_node}~$p$ in the horizontal list. If | |
| 156 | - the character represented by $p$ is considered as a Japanese | |
| 157 | - character, the font used at $p$ is replaced by the value of | |
| 158 | - |\ltj@curjfnt|, an attribute for `the current Japanese font' | |
| 159 | - at~$p$. | |
| 160 | - | |
| 161 | -Furthermore, the subtype of $p$ is subtracted by 1 to suppress | |
| 162 | - hyphenation around $p$ by \LuaTeX, because later processes of | |
| 163 | - \LuaTeX-ja take care of all things about Japanese characters. | |
| 164 | - | |
| 165 | -\item In |pre_linebreak_filter| and |hpack_filter| callbacks: | |
| 166 | - | |
| 167 | -\begin{enumerate} | |
| 168 | -\item \LuaTeX-ja has its own stack system, and the current horizontal | |
| 169 | - list is traversed in this stage to determine what the level of | |
| 170 | - \LuaTeX-ja's internal stack at the end of the list is. We will | |
| 171 | - discuss it in Subsection~\ref{ssec-stack}. | |
| 172 | - | |
| 173 | -\item In this stage, \LuaTeX-ja inserts glues/kerns for Japanese | |
| 174 | - typesetting in the list. This is the core routine of \LuaTeX-ja. | |
| 175 | - We will discuss it in Subsections | |
| 176 | - \ref{ssec-jglue}~and~\ref{ssec-jspec} . | |
| 177 | - | |
| 178 | -\item To make a match between a metric and a real font, sometimes | |
| 179 | - adjustument of the position of (Japanese) glyphs are performed. | |
| 180 | - We will discuss it in Subsection~\ref{ssec-width}. | |
| 181 | -\end{enumerate} | |
| 182 | -\item In the |mlist_to_hlist| callback: treatment of Japanese characters | |
| 183 | - in math formulas. This stage is similar to adjustment of the | |
| 184 | - position of glyphs (see above), so we omit to describe this stage | |
| 185 | - from this paper. | |
| 186 | -\end{itemize} | |
| 187 | - | |
| 188 | -In this paper, a \emph{alphabetic character} means a non-Japanese | |
| 189 | -character. Similarly, we use the word an \emph{alphabetic font} as the | |
| 190 | -counterpart of a jJpanese font. | |
| 191 | - | |
| 192 | -\subsection{Contents of this paper} | |
| 193 | -Here we describe the contents of the rest of this paper briefly. In | |
| 194 | -Section~\ref{sec:differences_with_ptex}, we describe major differences | |
| 195 | -between \pTeX\ and \LuaTeX-ja. The next section, | |
| 196 | -Section~\ref{sec:distinction_of_characters}, is concentrated on a | |
| 197 | -problem how we distinguish between Japanese characters and alphabetic | |
| 198 | -characters. In Section~\ref{sec:current_status}, we show current | |
| 199 | -development status of the package. Finally, in | |
| 200 | -Section~\ref{sec:implementation}, we describe some internal routines of | |
| 201 | -\LuaTeX-ja. | |
| 202 | - | |
| 203 | -\subsection{General information of the project} | |
| 204 | -This \LuaTeX-ja project is hosted by SourceForge.jp. The official wiki | |
| 205 | -is located on | |
| 206 | -\url{http://sourceforge.jp/projects/luatex-ja/wiki/}. There is | |
| 207 | -no stable version on October 22, 2011, however a set of developer sources can be | |
| 208 | -obtained from the git repository. Members of the project team are as follows | |
| 209 | -(in random order): Hironori Kitagawa, Kazuki Maeda, Takayuki Yato, | |
| 210 | -Yusuke Kuroki, Noriyuki Abe, Munehiro Yamamoto, Tomoaki Honda, | |
| 211 | -and~Shuzaburo Saito. | |
| 212 | - | |
| 213 | - | |
| 214 | -\section{Major differences with \pTeX} | |
| 215 | -\label{sec:differences_with_ptex} | |
| 216 | -In this section, we explain several major differences between \pTeX\ | |
| 217 | -and our \LuaTeX-ja. For general information of Japanese typesetting and the | |
| 218 | -overview of \pTeX, please see Okumura~\cite{ptexjp}. | |
| 219 | - | |
| 220 | - | |
| 221 | -\subsection{Names of control sequences} | |
| 222 | -\label{ssec-csname} Because \pTeX\ is an engine modification of Knuth's | |
| 223 | -original \TeX82 engine, some of the additional primitives take a form that is | |
| 224 | -very difficult to be simulated by a macro. For example, an additional | |
| 225 | -primitive |\prebreakpenalty|$\langle\hbox{\it | |
| 226 | -char\_code}\rangle$|[=]|$\langle\hbox{\it penalty}\rangle$ in \pTeX\ | |
| 227 | -sets the amount of penalty inserted before a character whose code is | |
| 228 | -$\langle\hbox{\it char\_code}\rangle$ to $\langle\hbox{\it | |
| 229 | -penalty}\rangle$, and this form |\prebreakpenalty|$\langle\hbox{\it | |
| 230 | -char\_code}\rangle$ can be also used for retrieving the value. | |
| 231 | - | |
| 232 | -Moreover, there are some internal parameters of \pTeX\ which values of them at the end of a | |
| 233 | -horizontal box or that of a paragraph are valid in whole box or | |
| 234 | -paragraph. However, the implementation of these parameters in | |
| 235 | -\LuaTeX-ja is not so easy; we will discuss it in Subsection~\ref{ssec-stack}. | |
| 236 | - | |
| 237 | -From above two~problems discussed above, the assignment and retrieval | |
| 238 | -of most parameters in \LuaTeX-ja are summarized into the following | |
| 239 | -three~control sequences: | |
| 240 | -\begin{itemize} | |
| 241 | -\item |\ltjsetparameter{|$\langle\hbox{\it | |
| 242 | - name}\rangle$|=|$\langle\hbox{\it value}\rangle$|,...}|: for local | |
| 243 | - assignment. | |
| 244 | -\item |\ltjglobalsetparameter|: for global assignment. Note that these two control | |
| 245 | - sequences obey the value of |\globaldefs| primitive. | |
| 246 | -\item |\ltjgetparameter{|$\langle\hbox{\it | |
| 247 | - name}\rangle$|}[{|$\langle\hbox{\it optional | |
| 248 | - argument}\rangle$|}]|: for retrieval. The returned value is always | |
| 249 | - a string. | |
| 250 | -\end{itemize} | |
| 251 | - | |
| 252 | -\subsection{Line-break after a Japanese character} | |
| 253 | -\label{ssec-line} | |
| 254 | - | |
| 255 | -Japanese texts can break lines almost everywhere, in contrast with | |
| 256 | -alphabetic texts can break lines only between words (or use | |
| 257 | -hyphenation). Hence, \pTeX's input processor is modified so that a | |
| 258 | -line-break after a Japanese character doesn't emit a space. However, | |
| 259 | -there is no way to customize the input processor of \LuaTeX, other than | |
| 260 | -to hack its CWEB-source. All a macro package can do is to modify an input line before | |
| 261 | -when \LuaTeX\ begin to process it, inside the |process_input_buffer| | |
| 262 | -callback. | |
| 263 | - | |
| 264 | -Hence, in \LuaTeX-ja, a comment letter (we reserve U+FFFFF for this | |
| 265 | -purpose) will be appended to an input line, if this line ends with a Japanese | |
| 266 | -character.\footnote{Strictly speaking, it also requires that the catcode | |
| 267 | -of the end-line character is 5~(\emph{end-of-line}). This condition is | |
| 268 | -useful under the verbatim environment.} One might jump to a conclusion | |
| 269 | -that the treatment of a line-break by \pTeX\ and that of \LuaTeX-ja are | |
| 270 | -totally same, however they are different in the respect that \LuaTeX-ja's | |
| 271 | -judgement whether a comment letter will be appended the line is done | |
| 272 | -\emph{before} the line is actually processed by \LuaTeX. | |
| 273 | - | |
| 274 | -Figure~\ref{fig-linebreak} shows an example of this situation; the | |
| 275 | -command at the first line marks most of Japanese characters as | |
| 276 | -`non-Japanese characters'. In other words, from that command onward, the | |
| 277 | -letter `あ' will be treated as an alphabetic character by | |
| 278 | -\LuaTeX-ja. Then, it is natural to have a space between `あ' and `y' in | |
| 279 | -the output, where the actual output in the figure does not so. This is | |
| 280 | -because `あ' is considered a Japanese character by \LuaTeX-ja, | |
| 281 | -when \LuaTeX-ja does the decision whether U+FFFFF will be added to the | |
| 282 | -input line~2. | |
| 283 | - | |
| 284 | -\begin{figure} | |
| 285 | -\begin{LTXexample} | |
| 286 | -\font\x=IPAMincho \x | |
| 287 | -\ltjsetparameter{jacharrange={-6}}xあ | |
| 288 | -y | |
| 289 | -\end{LTXexample} | |
| 290 | -\caption{A notable sample showing the treatment of a line-break after a | |
| 291 | -Japanese character.}\label{fig-linebreak} | |
| 292 | -\end{figure} | |
| 293 | - | |
| 294 | -\subsection{Separation between `real' fonts and metrics} | |
| 295 | -\label{ssec-sepmet} | |
| 296 | - | |
| 297 | -Traditionally, most Japanese fonts used in typesetting are not | |
| 298 | -proportional, that is, most glyphs have same size (in most cases, | |
| 299 | -square-shaped). Hence, it is not rare that the contents of different | |
| 300 | -JFMs are essentially same, and only differ in their names. For example, | |
| 301 | -|min10.tfm| and |goth10.tfm|, which are JFMs shipped with \pTeX\ for | |
| 302 | -seriffed \emph{mincho} family and sans-seriffed \emph{gothic} family, | |
| 303 | -differ their |FAMILY| and |FACE| only. Moreover, |jis.tfm| and | |
| 304 | -|jisg.tfm|, which is included in the \emph{jis} font metric, which is | |
| 305 | -used in \emph{jsclasses}~\cite{jsclasses} by Haruhiko Okumura (奥村晴彦), | |
| 306 | -are totally same as binary files. Considering this situation, we | |
| 307 | -decided to separate `real' fonts and metrics used for them in | |
| 308 | -\LuaTeX-ja. Typical declarations of Japanese fonts in the style of plain | |
| 309 | -\TeX\ are shown in Figure~\ref{fig-jfdef}. We would like to add several | |
| 310 | -remarks: | |
| 311 | -\begin{itemize} | |
| 312 | -\item A control sequence |\jfont| must be used for Japanese fonts, instead of |\font|. | |
| 313 | -\item \LuaTeX-ja automatically loads the \emph{luaotfload} package, so | |
| 314 | - \hbox{\tt file:} and \hbox{\tt name:} prefixes, and various font features can be | |
| 315 | - used as the first line in Figure~\ref{fig-jfdef}. | |
| 316 | -\item The |jfm| key specifies the metric for the font. In | |
| 317 | - Figure~\ref{fig-jfdef}, both fonts will use a metric stored in a | |
| 318 | - Lua script named |jfm-ujis.lua|. This metric is the standard | |
| 319 | - metric in \LuaTeX-ja, and is based on JFMs used in the \emph{otf} | |
| 320 | - package~\cite{otf}. | |
| 321 | -\item The \hbox{psft:} prefix can be used to specify name-only, non-embedded | |
| 322 | - fonts. When one display a pdf with these fonts, actual fonts which | |
| 323 | - will be used for them depend on a pdf reader. | |
| 324 | -\end{itemize} | |
| 325 | -The specification of a metric for \LuaTeX-ja is similar to that of a JFM | |
| 326 | -(see \cite{ptexjp}); characters are grouped into several classes, the | |
| 327 | -size information of characters are specified for each class, and | |
| 328 | -glue/kern insertions are specified for each pair of classes. Although | |
| 329 | -the author have not tried, it may be possible to develop a program that | |
| 330 | -`converts' a JFM to a metric for \LuaTeX-ja. \LuaTeX-ja offers three | |
| 331 | -metrics by default; |jfm-ujis.lua|, |jfm-jis.lua| based on the | |
| 332 | -\emph{jis} font metric, and |jfm-min.lua| based on old |min10.tfm|. | |
| 333 | - | |
| 334 | - Note that |-kern| in features | |
| 335 | -is important, because kerning information from a real font itself will | |
| 336 | -clash with glue/kern informations from the metric. | |
| 337 | - | |
| 338 | -\begin{figure} | |
| 339 | -\begin{verbatim} | |
| 340 | -\jfont\foo=file:ipam.ttf:jfm=ujis;script=latn;-kern;+jp04 at 12pt | |
| 341 | -\jfont\bar=psft:Ryumin-Light:jfm=ujis at 10pt | |
| 342 | -\end{verbatim} | |
| 343 | -\caption{Typical declarations of Japanese fonts.} | |
| 344 | -\label{fig-jfdef} | |
| 345 | -\end{figure} | |
| 346 | - | |
| 347 | -\subsection{Insertion of glues/kerns for Japanese typesetting: timing} | |
| 348 | -\label{ssec-jglue} | |
| 349 | - | |
| 350 | -As described in \cite{luatexref}, \LuaTeX's kerning and ligaturing | |
| 351 | -processes are totally different from those of \TeX82. \TeX82's process is | |
| 352 | -done just when a (sequence of) character is appended to the current | |
| 353 | -list. Thus we can interrupt this process by writing as | |
| 354 | -|f{}irm|. However, \LuaTeX's process is \emph{node-based}, that is, the | |
| 355 | -process will be done when a horizontal box or a paragraph is ended, so | |
| 356 | -|f{}irm| and |firm| yield same outputs under \LuaTeX. | |
| 357 | - | |
| 358 | -The situation for Japanese characters is more complicated. | |
| 359 | -Glues (and kerns) which are needed for Japanese | |
| 360 | -typesetting are divided into the following three categories: | |
| 361 | -\begin{itemize} | |
| 362 | -\item Glue (or kern) from the metric of Japanese fonts (\emph{JFM glue}, | |
| 363 | - for short). | |
| 364 | - | |
| 365 | -\item Default glue between a Japanese character and an alphabetic | |
| 366 | - character (\emph{xkanjiskip}, for short), usually 1/4 of | |
| 367 | - full-width (\emph{shibuaki}) with some stretch and shrink for | |
| 368 | - justifying each line. | |
| 369 | -\item Default glue between two consecutive Japanese characters | |
| 370 | - (\emph{kanjiskip}, for short). The main reason of this glue is to | |
| 371 | - enable breaking lines almost everywhere in Japanese texts. In most | |
| 372 | - cases, its natural width is zero, and some stretch/shrink for | |
| 373 | - justifying each line. | |
| 374 | -\end{itemize} | |
| 375 | -In \pTeX, these three kinds of glues are treated differently. A JFM glue | |
| 376 | -is inserted when a (sequence of) Japanese character is appended to the | |
| 377 | -current list, same as the case of alphabetic characters in \TeX82. This | |
| 378 | -means that one can interrupt the insertion process by saying |{}|. A | |
| 379 | -\emph{xkanjiskip} is inserted just before `hpack' or line-breaking of a | |
| 380 | -paragraph; this timing is somewhat similar to that of \LuaTeX's kerning | |
| 381 | -process. Finally, A \emph{kanjiskip} is not appeared as a node anywhere; | |
| 382 | -only appears implicitly in calculation of the width of a horizontal box, | |
| 383 | -that of breaking lines, and the actual output process to a DVI | |
| 384 | -file. These specifications made \pTeX's behavior very hard to | |
| 385 | -understand. | |
| 386 | - | |
| 387 | -\LuaTeX-ja inserts glues in all three categories simultaneously inside | |
| 388 | -|hpack_filter| and |pre_linebreak_filter| callbacks. The reasons of | |
| 389 | -this specification are to behave like alphabetic characters in \LuaTeX\ | |
| 390 | -(as described in the first paragraph), and to clarify the specification | |
| 391 | -for \LuaTeX-ja's process. | |
| 392 | - | |
| 393 | -\subsection{Insertion of glues/kerns for Japanese typesetting: specification} | |
| 394 | -\label{ssec-jspec} | |
| 395 | - | |
| 396 | -\begin{table} | |
| 397 | -\caption{Examples of differences between \pTeX\ and \LuaTeX-ja.} | |
| 398 | -\label{tab-jfmglue} | |
| 399 | -\begin{center} | |
| 400 | -\begin{tabular}{llllllll} | |
| 401 | -\toprule | |
| 402 | -&\multicolumn{1}{c}{(1)}&\multicolumn{1}{c}{(2)}&\multicolumn{1}{c}{(3)}&\multicolumn{1}{c}{(4)}\\ | |
| 403 | -Input &|あ】{}【〙\/〘| &|い』\/a| &|う）\hbox{}（| &|え］\special{}［|\\\midrule | |
| 404 | -\pTeX &あ】\hbox{}【〙\hbox{}〘&い』\/a &う）\hbox{}（ &え］\hbox{}［\\ | |
| 405 | -\LuaTeX-ja &あ】{}【〙\/〘 &い』\/a &う）\hbox{}（ &え］\special{}［\\ | |
| 406 | -\bottomrule | |
| 407 | -\end{tabular} | |
| 408 | -\end{center} | |
| 409 | -\end{table} | |
| 410 | - | |
| 411 | -\begin{figure} | |
| 412 | -\begin{center} | |
| 413 | -\fontsize{40}{40}\selectfont | |
| 414 | -\imagfm{\jstrut あ}% | |
| 415 | -\imagfm{\jstrut 】\inhibitglue}% | |
| 416 | -\imagfm{\jstrut\kern.5\zw}% | |
| 417 | -\imagfm{\jstrut\kern.5\zw}% | |
| 418 | -\imagfm{\jstrut\inhibitglue【}% | |
| 419 | -\imagfm{\jstrut 〙\inhibitglue}% | |
| 420 | -\imagfm{\jstrut\kern.5\zw}% | |
| 421 | -\imagfm{\jstrut\kern.5\zw}% | |
| 422 | -\imagfm{\jstrut\inhibitglue〘}% | |
| 423 | -\end{center} | |
| 424 | -\caption{Detail of (1) in Table~\ref{tab-jfmglue}.} | |
| 425 | -\label{fig-ptexjfm} | |
| 426 | -\end{figure} | |
| 427 | - | |
| 428 | -Now we will take a look inside the insertion process itself, and describe 4~points. | |
| 429 | - | |
| 430 | -\begin{description} | |
| 431 | -\item[Ignored Nodes] | |
| 432 | -As noted in the previous subsection, the insertion process in \pTeX\ can | |
| 433 | - be interrupted by saying |{}| or anything else.\footnote{This | |
| 434 | - is why some tricks like \texttt{ちょ\char`\{\char`\}っと} for | |
| 435 | - \texttt{min10.tfm} and other `old' JFMs work.} This leads the | |
| 436 | - second row in Table~\ref{tab-jfmglue}, or | |
| 437 | - Figure~\ref{fig-ptexjfm}. Here `the process is interrupted' | |
| 438 | - means that \pTeX\ does not think the letter `】\inhibitglue' | |
| 439 | - is followed by `\inhibitglue【', hence two half-width glues | |
| 440 | - are inserted between `】\inhibitglue' and `\inhibitglue【', | |
| 441 | - where the left one is from `】\inhibitglue' and the right one | |
| 442 | - is from `\inhibitglue【'. | |
| 443 | - | |
| 444 | - On the other hand, in \LuaTeX-ja, the process is done inside | |
| 445 | - |hpack_filter| and |pre_linebreak_filter| callbacks. Hence, | |
| 446 | - \emph{anything that does not make any node will be | |
| 447 | - ignored}\ in \LuaTeX-ja, as shown in (1) in | |
| 448 | - Table~\ref{tab-jfmglue}. \LuaTeX-ja also ignores any nodes | |
| 449 | - which does not make any contribution to current horizontal | |
| 450 | - list---\emph{ins\_node}, \emph{adjust\_node}, | |
| 451 | - \emph{mark\_node}, \emph{whatsit\_node} and | |
| 452 | - \emph{penalty\_node}---, as shown in (4). | |
| 453 | - | |
| 454 | - | |
| 455 | -By the way, around a \emph{glyph\_node} $p$ there may be some nodes | |
| 456 | - attached to $p$. These are an accent and kerns for | |
| 457 | - moving it to the right place, and a kern from the italic | |
| 458 | - correction\footnote{\TeX82 (and \LuaTeX) does not distinguish | |
| 459 | - between explicit kern and a kern for italic correction. To | |
| 460 | - distinguish them, an additional subtype for a kern is introduced | |
| 461 | - in \pTeX. On the other hand, \LuaTeX-ja uses an additional attribute and | |
| 462 | - redefines \texttt{\char`\\/} to set this attribute.} for $p$. It is natural that | |
| 463 | - these attachments should be ignored inside the process. Hence | |
| 464 | - \LuaTeX-ja takes this approach, as the latest version of | |
| 465 | - \pTeX\ (p3.2). This explains (2) in the figure. | |
| 466 | - | |
| 467 | -Summerizing above, one should put an empty horizontal box |\hbox{}| to | |
| 468 | - where he wants to interrupt the insertion process in | |
| 469 | - \LuaTeX-ja as (3) in the figure. | |
| 470 | - | |
| 471 | -\item[Fonts with the Same Metric] | |
| 472 | -Recall that \LuaTeX-ja separated `real' fonts and metrics, as in Subsection~\ref{ssec-sepmet}. | |
| 473 | -Consider the following input, where all Japanese fonts use same metric | |
| 474 | - (in \LuaTeX-ja), and |\gt| selects \emph{gothic} family for | |
| 475 | - the current Japanese font family: | |
| 476 | -\begin{quote} | |
| 477 | -\begin{verbatim} | |
| 478 | -明朝）\gt （ゴシック | |
| 479 | -\end{verbatim} | |
| 480 | -\end{quote} | |
| 481 | -If the above input is processed by \pTeX, because the insertion process is | |
| 482 | - interrupt by |\gt|, the result looks like | |
| 483 | -\begin{quote} | |
| 484 | -\mc 明朝）\hbox{}\gt （ゴシック | |
| 485 | -\end{quote} | |
| 486 | -However this seems to be unnatural, since two Japanese fonts in the | |
| 487 | - output use the same metric, i.e.,~the same | |
| 488 | - typesetting rule. Hence, we decided that Japanese fonts with | |
| 489 | - the same metric are treated as one font in the insertion | |
| 490 | - process of \LuaTeX-ja. Thus, the output from the above input | |
| 491 | - in \LuaTeX-ja looks like: | |
| 492 | -\begin{quote} | |
| 493 | -\mc 明朝）\gt （ゴシック | |
| 494 | -\end{quote} | |
| 495 | -One might have the situation that this default behavior is not | |
| 496 | - suitable. \LuaTeX-ja offers a way to handle this situation, but | |
| 497 | - we leave it to the manual~\cite{man}. | |
| 498 | - | |
| 499 | -\item[Fonts with Different Metrics] | |
| 500 | -In the case where two consecutive Japanese characters use different metrics and/or | |
| 501 | - different size is similar. Consider the following input where | |
| 502 | - the \emph{mincho} family and the \emph{gothic} family use | |
| 503 | - different metrics: | |
| 504 | -\begin{quote} | |
| 505 | -\begin{verbatim} | |
| 506 | -漢）\gt （漢）\large （大 | |
| 507 | -\end{verbatim} | |
| 508 | -\end{quote} | |
| 509 | -As the previous paragraph, this input yields the following, by \pTeX: | |
| 510 | -\begin{quote} | |
| 511 | -\mc 漢）\hbox{}\gt （漢）\hbox{}\large （大 | |
| 512 | -\end{quote} | |
| 513 | -We thought that amounts of spaces between parentheses in above output | |
| 514 | - are too much. Hence we changed the default behavior of | |
| 515 | - \LuaTeX-ja, so that the amount of a glue between two Japanese | |
| 516 | - characters with different metrics is the \emph{average} of a glue | |
| 517 | - from the left character and that from the right | |
| 518 | - character. For example, Figure~\ref{fig-diffmet} shows the | |
| 519 | - output from above input. The width of glue indicated `(1)' is | |
| 520 | - $(a/2 + a/2)/2 = 0.5a$, and the width of glue indicated `(2)' | |
| 521 | - is $(a/2 + 1.2a/2)/2 = 0.55a$. This default behavior can be | |
| 522 | - changed by \textsf{diffrentmet} parameter of \LuaTeX-ja. | |
| 523 | - | |
| 524 | -\begin{figure} | |
| 525 | -\begin{center} | |
| 526 | -\fontsize{40}{40}\selectfont | |
| 527 | -\imagfm{\jstrut\smash{% | |
| 528 | - \vtop{\lineskiplimit=\maxdimen\lineskip2pt\halign{#\cr漢\cr | |
| 529 | - \small\vrule height .5ex depth .5ex\hrulefill\ \lower.5ex\hbox{$a$}\ | |
| 530 | - \hrulefill\vrule height .5ex depth .5ex\cr}}}}% | |
| 531 | -\imagfm{\jstrut ）\inhibitglue}% | |
| 532 | -\hbox to .5\zw{\hss\normalsize (1)\hss}% | |
| 533 | -\imagfm{\jstrut\inhibitglue\gt （}% | |
| 534 | -\imagfm{\jstrut\gt 漢}% | |
| 535 | -\imagfm{\jstrut\gt ）\inhibitglue}% | |
| 536 | -\hbox to .55\zw{\hss\normalsize (2)\hss}% | |
| 537 | -\imagfm{\fontsize{48}{48}\selectfont\jstrut\gt\inhibitglue （}% | |
| 538 | -\imagfm{\fontsize{48}{48}\selectfont\jstrut\smash{% | |
| 539 | - \vtop{\lineskiplimit=\maxdimen\lineskip2pt\halign{#\cr\gt 大\cr | |
| 540 | - \small\vrule height .5ex depth .5ex\hrulefill\ \lower.5ex\hbox{$1.2a$}\ | |
| 541 | - \hrulefill\vrule height .5ex depth .5ex\cr}}}} | |
| 542 | -\end{center} | |
| 543 | -\caption{Fonts with different metrics.} | |
| 544 | -\label{fig-diffmet} | |
| 545 | -\end{figure} | |
| 546 | - | |
| 547 | -\item[\emph{kanjiskip} and \emph{xkanjiskip}] | |
| 548 | -In \pTeX, the value of \emph{xkanjiskip} is controlled by a skip named | |
| 549 | - |\xkanjiskip|. A well-known defect of this implementation is | |
| 550 | - that the value of \emph{xkanjiskip} is not connected with the | |
| 551 | - size of the currnt Japanese font. It seems that |EXTRASPACE|, | |
| 552 | - |EXTRASTRETCH|, |EXTRASHRINK| parameters in a JFM are | |
| 553 | - reserved for specifying the default value of | |
| 554 | - \emph{xkanjiskip} in a unit of the design size, but \pTeX\ | |
| 555 | - did not use these parameters, actually. | |
| 556 | - | |
| 557 | -Considering this situation of p\TeX, \LuaTeX-ja can use the value of | |
| 558 | - \emph{xkanjiskip} that specified in a metric. If the value of | |
| 559 | - \emph{xkanjiskip} on user side (this is the value of | |
| 560 | - \textsf{xkanjiskip} parameter of |\ltjsetparameter|) is | |
| 561 | - |\maxdimen|, then \LuaTeX-ja use the specification from | |
| 562 | - the current used metric as the actual value of | |
| 563 | - \emph{xkanjiskip}. This description also applies for \emph{kanjiskip}. | |
| 564 | -\end{description} | |
| 565 | - | |
| 566 | -\section{Distinction of characters} | |
| 567 | -\label{sec:distinction_of_characters} Since \LuaTeX\ can handle Unicode | |
| 568 | -characters natively, it is a major problem that how we distinguish | |
| 569 | -Japanese characters and alphabetic characters. For example, the | |
| 570 | -multiplication sign (U+00D7) exists both in ISO-8859-1 (hence in Latin-1 | |
| 571 | -Supplement in Unicode) and in the basic Japanese character set | |
| 572 | -JIS~X~0208. It is not desirable that this character is always treated as | |
| 573 | -an alphabetic character, because this symbol is often used in the sense | |
| 574 | -of `negative' in Japan. | |
| 575 | - | |
| 576 | -\subsection{Character ranges} | |
| 577 | -Before we describe the approach taken is \LuaTeX-ja, we review the | |
| 578 | -approach taken by u\pTeX. u\pTeX\ extends the |\kcatcode| primitive in | |
| 579 | -\pTeX, to use this primitive for setting how a character is treated | |
| 580 | -among alphabetic characters~(15), \emph{kanji}~(16), \emph{kana}~(17), | |
| 581 | -\emph{kanji}, \emph{Hangul}~(17), or~\emph{other CJK characters}~(18). | |
| 582 | -The assignment to |\kcatcode| can be done by a Unicode | |
| 583 | -block.\footnote{There are some exceptions. For example, U+FF00--FFEF | |
| 584 | -(Halfwidth and Fullwidth Forms) are divided into three blocks in recent | |
| 585 | -u\pTeX.} | |
| 586 | - | |
| 587 | -\LuaTeX-ja adopted a different approach. There are many Unicode blocks | |
| 588 | - in Basic Multilingual Plane which are not included in | |
| 589 | - Japanese fonts, therefore it is inconvenient if we process by a Unicode | |
| 590 | - block. Furthermore, JIS~X~0208 are not just union of Unicode | |
| 591 | - blocks; for example, the intersection of JIS~X~0208 and | |
| 592 | - Latin-1 Supplement is shown in | |
| 593 | - Table~\ref{tab-inter}. Considering these two points, to | |
| 594 | - customize the range of Japanese characters in \LuaTeX-ja, one | |
| 595 | - has to define ranges of character codes in his source in advance. | |
| 596 | - | |
| 597 | - | |
| 598 | -\begin{table} | |
| 599 | -\caption{Intersection of JIS~X~0208 and Latin-1 Supplement.} | |
| 600 | -\label{tab-inter} | |
| 601 | -\begin{center} | |
| 602 | -\begin{tabular}{llll} | |
| 603 | -\ltjjachar"A7 (U+00A7),& | |
| 604 | -\ltjjachar"A8 (U+00A8),& | |
| 605 | -\ltjjachar"B0 (U+00B0),& | |
| 606 | -\ltjjachar"B1 (U+00B1),\\ | |
| 607 | -\ltjjachar"B4 (U+00B4),& | |
| 608 | -\ltjjachar"B6 (U+00B6),& | |
| 609 | -\ltjjachar"D7 (U+00D7),& | |
| 610 | -\ltjjachar"F7 (U+00F7) | |
| 611 | -\end{tabular} | |
| 612 | -\end{center} | |
| 613 | -\end{table} | |
| 614 | - | |
| 615 | -%%Example... | |
| 616 | - | |
| 617 | -We note that \LuaTeX-ja offers two additional control sequences, | |
| 618 | - |\ltjjachar| and |\ltjalchar|. They are similar to |\char| | |
| 619 | - primitive, however |\ltjjachar| always yields a Japanese character, provided that | |
| 620 | - the argument is more than or equal to 128, and |\ltjalchar| always | |
| 621 | - yields an alphabetic character, regardless of the argument. | |
| 622 | - | |
| 623 | -\subsection{Default setting of ranges} | |
| 624 | -Patches for plain \TeX\ and \LaTeXe\ of \LuaTeX-ja predefine 8~character | |
| 625 | -ranges, as shown in Table~\ref{tab-chrrng}. Almost of these ranges are | |
| 626 | -just the union of Unicode blocks, and determined from the Adobe-Japan1-6 | |
| 627 | -character collection~\cite{aj16}, and JIS~X~0208. Among these 8~ranges, | |
| 628 | -the ranges~2, 3, 6, 7, and~8 are considered ranges of Japanese | |
| 629 | -characters, and others are considered ranges of alphabetic | |
| 630 | -characters\footnote{Note that ranges 3~and~8 are considered ranges of | |
| 631 | -alphabetic characters in this paper.}. We remark on ranges 2~and~8: | |
| 632 | -\begin{description} | |
| 633 | -\item[The range~2] | |
| 634 | -JIS~X~0208 includes Greek letters and Cyrillic letters, however, these | |
| 635 | - letters cannot be used for typesetting Greek or Russian, of | |
| 636 | - course. Hence it is reasonable that Greek letters and | |
| 637 | - Cyrillic consist another character range. | |
| 638 | -\item[The range~8] | |
| 639 | -If one want to use 8-bit TFMs, such as T1 or TS1 encodings, he should | |
| 640 | - mark this range~8 as a range of alphabetic characters by | |
| 641 | -\begin{quote} | |
| 642 | -|\ltjsetparameter{jacharrange={-8}}| | |
| 643 | -\end{quote} | |
| 644 | -This is because some 8-bit TFMs have a glyph in this range; for example, | |
| 645 | - the character `\OE' is located at |"D7| in the T1 encoding. %" | |
| 646 | -\end{description} | |
| 647 | - | |
| 648 | - | |
| 649 | -\begin{table} | |
| 650 | -\caption{Predefined ranges in \LuaTeX-ja.} | |
| 651 | -\label{tab-chrrng} | |
| 652 | -\begin{center} | |
| 653 | -\begin{tabular}{@{\bf}rl} | |
| 654 | -1&(Additional) Latin characters which are not belonged in the range~8.\\ | |
| 655 | -2&Greek and Cyrillic letters.\\ | |
| 656 | -3&Punctuations and miscellaneous symbols.\\ | |
| 657 | -4&Unicode blocks which does not intersect with Adobe-Japan1-6.\\ | |
| 658 | -5&Surrogates and supplementary private use Areas.\\ | |
| 659 | -6&Characters used in Japanese typesetting.\\ | |
| 660 | -7&Characters possibly used in CJK typesetting, but not in Japanese.\\ | |
| 661 | -8&Characters in Table~\ref{tab-inter}. | |
| 662 | -\end{tabular} | |
| 663 | -\end{center} | |
| 664 | -\end{table} | |
| 665 | - | |
| 666 | -\subsection{Control sequences producing Unicode characters} | |
| 667 | -\label{ssec-unichar} | |
| 668 | - | |
| 669 | -The \emph{fontspec} package\footnote{Preciously saying, it is the | |
| 670 | -\emph{xunicode} package, originally a package for \XeTeX and | |
| 671 | -automatically loaded by the \emph{fontspec} package.} offers various | |
| 672 | -control sequences that produce Unicode characters. However, these | |
| 673 | -control sequences as it stands cannot work correctly with the default | |
| 674 | -range setting of \LuaTeX-ja. For example, |\textquotedblleft| is just | |
| 675 | -an abbreviation of |\char"201C\relax|, and the character U+201C (LEFT %" | |
| 676 | -DOUBLE QUOTATION MARK) is treated as an Japanese character, because it | |
| 677 | -belongs to the range~3. This problem is resolved by using |\ltjalchar| | |
| 678 | -instead of the |\char| primitive. It is included in an optional package | |
| 679 | -named \texttt{luatexja-\penalty0fontspec.sty}. Figure~\ref{fig-unitxt} | |
| 680 | -shows several ways o typeset a character , both as a Japanese character | |
| 681 | -and as as an alphabetic characters. | |
| 682 | - | |
| 683 | -\begin{figure} | |
| 684 | -\begin{LTXexample} | |
| 685 | -×, \char`×, % depend on range setting | |
| 686 | -\ltjalchar`×, % alphabetic char | |
| 687 | -\ltjjachar`×, % Japanese char | |
| 688 | -\texttimes % alph. char (by fontspec) | |
| 689 | -\end{LTXexample} | |
| 690 | -\caption{Control sequences producing a Unicode character.} | |
| 691 | -\label{fig-unitxt} | |
| 692 | -\end{figure} | |
| 693 | - | |
| 694 | -The situation looks similar in math formulas, but in fact it differs. | |
| 695 | -Each control sequence that represents an ordinary symbol defined by the | |
| 696 | -\emph{unicode-math} package is just synonym of a character. For example, | |
| 697 | -the meaning of |\otimes| is just the character U+2297 (CIRCLED TIMES), | |
| 698 | -which is included in the range~3. However, it is difficult to define a | |
| 699 | -control sequence like |\ltjalUmathchar| as a counterpart of | |
| 700 | -|\Umathchar|, since an input like `|\sum^\ltjalUmathchar ...|' has to be | |
| 701 | -permitted. | |
| 702 | - | |
| 703 | -However, we couldn't develop a satisfactory solution to this problem in | |
| 704 | -time for this paper, due to a lack of time. We are just testing a | |
| 705 | -solution below: | |
| 706 | -\begin{itemize} | |
| 707 | -\item \LuaTeX-ja has a list of character codes which will be always reated as | |
| 708 | - alphabetic characters in math mode. Considering 8-bit TFMs for | |
| 709 | - math symbols, this list includes natural numbers between |"80| and | |
| 710 | - |"FF| by default. | |
| 711 | -\item Redefine internal commands defined in the \emph{unicode-math} | |
| 712 | - package so that | |
| 713 | -codes of characters which are mentioned in the \emph{unicode-math} | |
| 714 | - package will be included in the list. | |
| 715 | -\end{itemize} | |
| 716 | - | |
| 717 | - | |
| 718 | -We would like to extend treatments described in this subsection to 8-bit | |
| 719 | -font encodings, but we leave it to further development too. | |
| 720 | - | |
| 721 | -\section{Current status of development} | |
| 722 | -\label{sec:current_status} | |
| 723 | -At the moment, \LuaTeX-ja can be used under plain \TeX, and under | |
| 724 | -\LaTeXe. Generally speaking, one only has to read |luatexja.sty|, by | |
| 725 | -|\input| command or |\usepackage| (in~\LaTeXe), if you merely want to | |
| 726 | -typeset Japanese characters. We look more detail by parts. | |
| 727 | - | |
| 728 | -\subsection{`Engine extension'} | |
| 729 | -The lowest part of \LuaTeX-ja corresponds the \pTeX\ extension as | |
| 730 | -\emph{an engine extension of \TeX}. We, the project menbers, think that | |
| 731 | -this part is almost done. There is one more feature of \LuaTeX-ja which | |
| 732 | -we are going to explain: | |
| 733 | - | |
| 734 | -\begin{description} | |
| 735 | -\item[Shifting Baseline] | |
| 736 | -In order to make a match between Japanese fonts and alphabetic fonts, | |
| 737 | - sometimes shifting the baseline of alphabetic characters may | |
| 738 | - be needed. \pTeX\ has a dimension |\ybaselineshift|, which | |
| 739 | - corresponds the amount of shifting down the baseline of alphabetic | |
| 740 | - characters. This is useful for Japanese-based documents, but | |
| 741 | - not for documents mainly in languages with alphabetic | |
| 742 | - characters. | |
| 743 | - | |
| 744 | -Hence, \LuaTeX-ja extends \pTeX's |\ybaselineshift| to Japanese | |
| 745 | - characters. Namely, \LuaTeX-ja offers two parameters, | |
| 746 | - \textsf{yjabaselineshift} and \textsf{yalbaselineshift}, for the | |
| 747 | - amount of shifting the baseline of Japanese characters and | |
| 748 | - that of alphabetic characters, respectively. | |
| 749 | -\begin{figure} | |
| 750 | -\begin{center} | |
| 751 | -\fontsize{40}{40}\selectfont\fboxsep0mm | |
| 752 | -\vrule width 0.9\textwidth height0.4pt depth0.4pt\kern-0.9\textwidth | |
| 753 | -\hbox to 0.9\linewidth{% | |
| 754 | -\hfil | |
| 755 | -\raise-10pt\imagfm{\jstrut 漢}% | |
| 756 | -\raise-10pt\imagfm{\jstrut 字}\hskip.25\zw% | |
| 757 | -\imagfm{p}% | |
| 758 | -\imagfm{h}% | |
| 759 | -\hfil\hfil | |
| 760 | -\imagfm{\jstrut 漢}% | |
| 761 | -\imagfm{\jstrut 字}\hskip.25\zw% | |
| 762 | -\raise-10pt\imagfm{p}% | |
| 763 | -\raise-10pt\imagfm{h}% | |
| 764 | -\hfil | |
| 765 | -} | |
| 766 | -\end{center} | |
| 767 | - | |
| 768 | -\caption{First example of shifting baseline.} | |
| 769 | -\label{fig-bls} | |
| 770 | -\end{figure} | |
| 771 | - | |
| 772 | -\begin{figure} | |
| 773 | -\begin{center} | |
| 774 | -\fontsize{30}{30}\selectfont\fboxsep0mm | |
| 775 | -\vrule width 0.9\textwidth height0.4pt depth0.4pt\kern-0.9\textwidth | |
| 776 | -\hbox to 0.9\linewidth{% | |
| 777 | -\hfil | |
| 778 | -\imagfm{a}% | |
| 779 | -\imagfm{b}\hskip.25\zw% | |
| 780 | -\imagfm{\jstrut 本}% | |
| 781 | -\imagfm{\jstrut 文}\hskip.33333\zw% | |
| 782 | -\raise3.514582pt\imagfm{\fontsize{20}{20}\selectfont\jstrut\inhibitglue （}% | |
| 783 | -\raise3.514582pt\imagfm{\fontsize{20}{20}\selectfont\jstrut 注}% | |
| 784 | -\raise3.514582pt\imagfm{\fontsize{20}{20}\selectfont\jstrut 釈}\hskip.1666667\zw% | |
| 785 | -\raise3.514582pt\imagfm{\fontsize{20}{20}\selectfont c}% | |
| 786 | -\raise3.514582pt\imagfm{\fontsize{20}{20}\selectfont o}% | |
| 787 | -\raise3.514582pt\imagfm{\fontsize{20}{20}\selectfont m}% | |
| 788 | -\raise3.514582pt\imagfm{\fontsize{20}{20}\selectfont m}% | |
| 789 | -\raise3.514582pt\imagfm{\fontsize{20}{20}\selectfont e}% | |
| 790 | -\raise3.514582pt\imagfm{\fontsize{20}{20}\selectfont n}% | |
| 791 | -\raise3.514582pt\imagfm{\fontsize{20}{20}\selectfont t}% | |
| 792 | -\raise3.514582pt\imagfm{\fontsize{20}{20}\selectfont\jstrut ）\inhibitglue}% | |
| 793 | -\hskip.33333\zw% | |
| 794 | -\imagfm{\jstrut 本}% | |
| 795 | -\imagfm{\jstrut 文}% | |
| 796 | -\hfil | |
| 797 | -} | |
| 798 | -\end{center} | |
| 799 | - | |
| 800 | -\caption{Second example of shifting baseline.} | |
| 801 | -\label{fig-small} | |
| 802 | -\end{figure} | |
| 803 | - | |
| 804 | -An example output is shown in Figure~\ref{fig-bls}. The left half is the | |
| 805 | - output when \textsf{yjabaselineshift} is positive, hence the | |
| 806 | - baseline of Japanese characters is shifted down. On the other | |
| 807 | - hand, the right half is the output when | |
| 808 | - \textsf{yalbaselineshift} is positive, hence the baseline of | |
| 809 | - alphabetic characters is shifted down. Figure~\ref{fig-small} | |
| 810 | - shows an intresting use of these parameters. | |
| 811 | - | |
| 812 | -\end{description} | |
| 813 | -Note that \LuaTeX-ja doesn't support vertical typesetting, \emph{tategaki}, for now. | |
| 814 | - | |
| 815 | -\subsection{Patches for plain \TeX\ and \LaTeXe} | |
| 816 | -\pTeX\ has a patch for plain \TeX, namely |ptex.tex|, that for \LaTeXe\ | |
| 817 | -macro (this patch and \LaTeXe\ consist \emph{p\LaTeXe}), and | |
| 818 | -|kinsoku.tex| which includes the default setting of \emph{kinsoku | |
| 819 | -shori}, the Japanese hyphenation. We ported them to \LuaTeX-ja, except | |
| 820 | -the codes related to vertical typesetting, because \LuaTeX-ja doesn't | |
| 821 | -support vertical typesetting yet. We remark one point related to the | |
| 822 | -porting: | |
| 823 | -\begin{description} | |
| 824 | - | |
| 825 | -\item[Behavior of\/ {\tt\char92fontfamily\/}] | |
| 826 | -The control sequence |\fontfamily| in p\LaTeXe\ changes the current alphabetic | |
| 827 | - font family and/or the current Japanese font family, | |
| 828 | - depending the argument. More concretely, | |
| 829 | - |\fontfamily{|$\langle\hbox{\it arg\/}\rangle$|}| changes the | |
| 830 | - current alphabetic font family to $\langle\hbox{\it | |
| 831 | - arg\/}\rangle$, if and only if one of the following | |
| 832 | - conditions are satisfied: | |
| 833 | -\begin{itemize} | |
| 834 | -\item An alphabetic font family named $\langle\hbox{\it arg\/}\rangle$ in | |
| 835 | - \emph{some} alphabetic encoding already defined in the document. | |
| 836 | -\item There exists an alphabetic encoding $\langle\hbox{\it | |
| 837 | - enc\/}\rangle$ already defined in the document such that a font | |
| 838 | - definition file $\langle\hbox{\it enc\/}\rangle\langle\hbox{\it | |
| 839 | - arg\/}\rangle$|.fd| (all lowercase) exists. | |
| 840 | -\end{itemize} | |
| 841 | -The same criterion is used for changing Japanese font family. | |
| 842 | - | |
| 843 | -To work this behavior well, a list of all (alphabetic) encodings defined | |
| 844 | - already in the document is needed. However, since \LuaTeX-ja | |
| 845 | - is loaded as a package, \LuaTeX-ja cannot have this list. | |
| 846 | - Hence \LuaTeX-ja adopted a different approach, namely | |
| 847 | - |\fontfamily{|$\langle\hbox{\it arg\/}\rangle$|}| changes the | |
| 848 | - current alphabetic font family to $\langle\hbox{\it | |
| 849 | - arg\/}\rangle$, if and only if: | |
| 850 | -\begin{itemize} | |
| 851 | -\item An alphabetic font family named $\langle\hbox{\it arg\/}\rangle$ | |
| 852 | - in the current alphabetic encoding $\langle\hbox{\it | |
| 853 | - enc\/}\rangle$ already defined in the document. | |
| 854 | -\item A font definition file $\langle\hbox{\it enc\/}\rangle\langle\hbox{\it | |
| 855 | - arg\/}\rangle$|.fd| (all lowercase) exists. | |
| 856 | -\end{itemize} | |
| 857 | - | |
| 858 | - | |
| 859 | -\end{description} | |
| 860 | - | |
| 861 | - | |
| 862 | - | |
| 863 | -\subsection{Classes for Japanese documents} | |
| 864 | -To produce `high-quality' Japanese documents, we need not only that | |
| 865 | -Japanese characters are correctly placed, but also class files for | |
| 866 | -Japanese documents. Two major families of classes are widely used in Japan: | |
| 867 | -\emph{jclasses} which is distributed with the official p\LaTeXe\ macros, | |
| 868 | -and \emph{jsclasses}. At the present, \LuaTeX-ja | |
| 869 | -simply contains their counterparts: \emph{ltjclasses} and | |
| 870 | -\emph{ltjsclasses}. However, the policy on classes is not determined | |
| 871 | -now, and we hope to have another family of classes which are useful for | |
| 872 | -commercial printing. In the author's opinion, \emph{ltjclasses} is | |
| 873 | -better to stay as an example of porting of class files for \pTeX\ to | |
| 874 | -\LuaTeX-ja. | |
| 875 | - | |
| 876 | -\subsection{Patches for packages} | |
| 877 | -Apart from patches for the \LaTeXe~kernel and classes for Japanese | |
| 878 | -documents, we need to make patches for several packages. At the present, | |
| 879 | -we considered the following packages, and made patches or porting for | |
| 880 | -the former two packages. | |
| 881 | - | |
| 882 | -\begin{description} | |
| 883 | -\item[The \emph{fontspec} package] The \emph{fontspec} package is built | |
| 884 | - on NFSS2, hence control sequences offered by the | |
| 885 | - \emph{fontspec} package, such as |\setmainfont|, are only | |
| 886 | - effective for alphabetic fonts if \LuaTeX-ja is loaded. | |
| 887 | - \texttt{luatexja-\penalty0fontspec.sty} (not automatically | |
| 888 | - loaded) offers these counterparts for Japanese fonts, with | |
| 889 | - additional `j' in the name of control sequences, such as | |
| 890 | - |\setmainjfont|. As described in | |
| 891 | - Subsection~\ref{ssec-unichar}, it also includes a patch for | |
| 892 | - control sequences producing Unicode characters. | |
| 893 | - | |
| 894 | -\item[The \emph{otf} package] | |
| 895 | -This package is widely used in \pTeX\ for typesetting characters which is | |
| 896 | -not in JIS~X~0208, and for using more than one weight in \emph{mincho} | |
| 897 | -and \emph{gothic} font families. Therefore \LuaTeX-ja supports features | |
| 898 | -in the \emph{otf} package, by loading \texttt{luatexja-\penalty0otf.sty} | |
| 899 | - manually. Note that characters by |\UTF{xxxx}| and | |
| 900 | - |\CID{xxxx}| are not appended to the current list as a | |
| 901 | - \emph{glyph\_node}, to avoid from callbacks by the | |
| 902 | - \emph{luaotfload} package. We have another remark; |\CID| | |
| 903 | - does not work with TrueType fonts, since |\CID| use the | |
| 904 | - conversion table between CID and the glyph order of the | |
| 905 | - current Japanese font. | |
| 906 | - | |
| 907 | -\item[The \emph{listings} package] | |
| 908 | -It is known for users of \pTeX\ that there is a patch |jlisting.sty| for | |
| 909 | - the \emph{listings} package, to use Japanese characters in | |
| 910 | - the |lstlisting| environment. Generally speaking, it also can | |
| 911 | - be used in \LuaTeX-ja. However, it seems to be that a | |
| 912 | - Japanese character after a space does not recieve any process | |
| 913 | - of the \emph{listings} package; this is inconvinient when we | |
| 914 | - use the \emph{showexpl} package. | |
| 915 | - | |
| 916 | -There is another way to use characters above 256 with the | |
| 917 | - \emph{listings} package (described in\cite{apl}). However, | |
| 918 | - this method is not suitable for Japanese, since the number of | |
| 919 | - Japanese characters is very large. We hope that the | |
| 920 | - \emph{listings} package will be able to handle all characters above | |
| 921 | - 256 without any patch, in the future. | |
| 922 | - | |
| 923 | - | |
| 924 | -\end{description} | |
| 925 | - | |
| 926 | - | |
| 927 | - | |
| 928 | -\section{Implementation} | |
| 929 | -\label{sec:implementation} | |
| 930 | -\subsection{Handling of Japanese fonts} | |
| 931 | -In \pTeX, there are three slots for maintaining current fonts, namely | |
| 932 | -|\font| for alphabetic fonts, |\jfont| for Japanese fonts (in horizontal | |
| 933 | -direction) and |\tfont| for Japanese fonts (in vertical direction). With | |
| 934 | -these slots, we can manage the current font for alphabetic characters | |
| 935 | -and that for Japanese characters separately in \pTeX. However, \LuaTeX\ | |
| 936 | -has only one slot for maintaining the current font, as \TeX82. This | |
| 937 | -situation leads a problem: how can we maintain the `current Japanese | |
| 938 | -font'? | |
| 939 | - | |
| 940 | -There are three approaches for this problem. One approach is to make a | |
| 941 | -mapping table from alphabetic fonts to corresponding Japanese fonts | |
| 942 | -(here we don't assume that NFSS2 is available). Another approach is | |
| 943 | -that we always use composite fonts with alphabetic fonts and Japanese | |
| 944 | -fonts. The third approach is that the information of the current | |
| 945 | -Japanese font is stored in an attribute. We adopted the third approach, | |
| 946 | -since \LuaTeX-ja is much affected by \pTeX\ as we noted in | |
| 947 | -Subsection~\ref{ssec-pol}. | |
| 948 | - | |
| 949 | -As in Figure~\ref{fig-jfdef}, \LuaTeX-ja uses |\jfont| for defining | |
| 950 | -Japanese font, as \pTeX. However, because the information of the current | |
| 951 | -Japanese font is stored into an attribute, control sequences defined by | |
| 952 | -|\jfont| (e.g.,~|\foo| and |\bar| in Figure~\ref{fig-jfdef}) is | |
| 953 | -not representing a font by the means of \TeX82. In other words, each of | |
| 954 | -these control sequences is just an assignment to an attribute, therefore | |
| 955 | -they cannot be an argument of |\the|, |\fontname|, nor |\textfont|. | |
| 956 | - | |
| 957 | - | |
| 958 | -Callbacks by the \emph{luaotfload} package, e.g.,~replacement of glyphs | |
| 959 | -according to font features, are executed just after `Examination of | |
| 960 | -Stack Level' (see Subsections \ref{ssec-over}~and~\ref{ssec-stack}). Note that calculation of | |
| 961 | -character classes for each Japanese character is done \emph{after} the | |
| 962 | -these callbacks for now. | |
| 963 | - | |
| 964 | -\subsection{Stack management} | |
| 965 | -\label{ssec-stack} | |
| 966 | - | |
| 967 | -As we noted in Subsection~\ref{ssec-csname}, parameters that the values | |
| 968 | -at the end of a horizontal box or that of a paragraph are valid in | |
| 969 | -whole box or paragraph, such as \emph{kanjiskip}, cannot be implemented | |
| 970 | -by internal integers or registers of other types in \TeX. We explain it | |
| 971 | -in this subsection. | |
| 972 | - | |
| 973 | -\begin{figure} | |
| 974 | -\begin{lstlisting} | |
| 975 | -void package(int c) | |
| 976 | -{ | |
| 977 | - ... | |
| 978 | - d = box_max_depth; | |
| 979 | - unsave(); | |
| 980 | - save_ptr -= 4; | |
| 981 | - if (cur_list.mode_field == -hmode) { | |
| 982 | - cur_box = filtered_hpack(cur_list.head_field, | |
| 983 | - cur_list.tail_field, saved_value(1), | |
| 984 | - saved_level(1), grp, saved_level(2)); | |
| 985 | - subtype(cur_box) = HLIST_SUBTYPE_HBOX; | |
| 986 | - } else { | |
| 987 | -\end{lstlisting} | |
| 988 | -\caption{An extract of a CWEB-source \texttt{tex/packaging.w} of \LuaTeX.} | |
| 989 | -\label{fig-ltsrc} | |
| 990 | -\end{figure} | |
| 991 | - | |
| 992 | -Figure~\ref{fig-ltsrc} is an extract of a CWEB-source | |
| 993 | -\texttt{tex/packaging.w} of \LuaTeX\ (SVN revision 4358). This function | |
| 994 | -is called just when an explicit |\hbox{...}| or |\vbox{...}| is ended, and | |
| 995 | -the function |filtered_hpack()| is where the |hpack_filter| and then the | |
| 996 | -actual `hpack' process are performed. Notice that the |unsave()| | |
| 997 | -function is called before |filtered_hpack()|. This is the problem; | |
| 998 | -because of |unsave()|, we can retrive only the values of registers | |
| 999 | -\emph{outside} the box, even in the |hpack_filter| callback. | |
| 1000 | - | |
| 1001 | -To cope with this problem, \LuaTeX-ja has its own stack system, based on | |
| 1002 | -Lua codes in \cite{stack-mail}. Furthermore, \emph{whatsit} nodes whose | |
| 1003 | -\emph{user\_id} is 30112 (\emph{stack\_node}, for short) will be | |
| 1004 | -appended to the current horizontal list each time the current stack | |
| 1005 | -level is incremented, and their values are the values of | |
| 1006 | -|\currentgrouplevel| at that time. In the beginning of the |hpack_filter| | |
| 1007 | -callback, the list in question is traversed to determine whether the | |
| 1008 | -stack level at the end of the list and that outside the box coincides. | |
| 1009 | - | |
| 1010 | -Let $x$ be the value of |\currentgrouplevel|, and $y$ be the current | |
| 1011 | -stack level, both inside the |hpack_filter| callback, i.e.,~outside a | |
| 1012 | -horizontal box. Consider a list which represents the content of the box, | |
| 1013 | -then we have: | |
| 1014 | -\begin{itemize} | |
| 1015 | -\item A \emph{stack\_node} whose value is $x+1$ (because all materials in | |
| 1016 | - the box are included in a group |\hbox{...}|, the value is at | |
| 1017 | - least $x+1$) in the list represents an assignment related to the | |
| 1018 | - stack system in just top-level of the list, like | |
| 1019 | -\begin{quote} | |
| 1020 | -\begin{verbatim} | |
| 1021 | -\hbox{...(assignment)...} | |
| 1022 | -\end{verbatim} | |
| 1023 | -\end{quote} | |
| 1024 | -In this case, the current stack level is incremented to $y+1$ after the assignment. | |
| 1025 | -\item A \emph{stack\_node} whose value is more than $x+1$ in the list represents | |
| 1026 | -an assignment inside another group contained in the box. For example, | |
| 1027 | - the following input creates | |
| 1028 | -a \emph{stack\_node} whose value is $x+3=(x+1)+2$: | |
| 1029 | -\begin{quote} | |
| 1030 | -\begin{verbatim} | |
| 1031 | -\hbox{...{...{...(assignment)}...}...} | |
| 1032 | -\end{verbatim} | |
| 1033 | -\end{quote} | |
| 1034 | -\end{itemize} | |
| 1035 | -Thus, we can conclude that the stack level at the end of the list is | |
| 1036 | -$y+1$, if and only if there is a \emph{stack\_node} whose value is | |
| 1037 | -$x+1$. Otherwise, the stack level is just $y$. | |
| 1038 | - | |
| 1039 | -\subsection{Adjustment of the position of Japanese characters} | |
| 1040 | -\label{ssec-width} | |
| 1041 | - | |
| 1042 | -The size of a glyph specified in a metric and that of a real font | |
| 1043 | -usually differ. For example, the letter `\inhibitglue【' is half-width | |
| 1044 | -in |jfm-ujis.lua| or |jis.tfm|, while this letter is full-width like `【' | |
| 1045 | -in most TrueType fonts used in Japanese typesetting, such as | |
| 1046 | -IPA~Mincho. Hence the adjustment of position of such glyphs is | |
| 1047 | -needed. In the context of \pTeX, this process was performed using virtual fonts. | |
| 1048 | - | |
| 1049 | -On the other hand, Lua\TeX-ja does the adjustment by encapsuling a glyph | |
| 1050 | -into a horizontal box. There are two main reasons why we adopted this | |
| 1051 | -method; one is that we feared Lua codes for coexisting with callbacks by | |
| 1052 | -the |luaotfload| package would be large if we use virtual fonts, and the | |
| 1053 | -other is to cope with shifting of the baseline of characters at the | |
| 1054 | -same time. | |
| 1055 | - | |
| 1056 | -\begin{figure} | |
| 1057 | -\begin{center}\unitlength=9pt\small | |
| 1058 | -\begin{picture}(15,12)(-1,-3) | |
| 1059 | - | |
| 1060 | -\color{grayx}% real glyph | |
| 1061 | -\put(-1,-1.5){\vrule width 6\unitlength height 7\unitlength depth 2.5\unitlength} | |
| 1062 | - | |
| 1063 | -\color{black}% real glyph :step1 | |
| 1064 | -\thicklines | |
| 1065 | -\put(-1,-1.5){\line(0,1){7}\line(0,-1){2.5}} | |
| 1066 | -\put(5,-1.5){\line(0,1){7}\line(0,-1){2.5}} | |
| 1067 | -\put(-1,5.5){\line(1,0){6}} | |
| 1068 | -\put(-1,-4){\line(1,0){6}} | |
| 1069 | -\put(-1,0){\makebox(0,0)[r]{\strut$R$\,}} | |
| 1070 | - | |
| 1071 | -\thicklines | |
| 1072 | -\put(0,0){\vector(0,1){9}\line(0,-1){3}\vector(1,0){12}} | |
| 1073 | -\put(12,9){\makebox(0,0)[rt]{\strut$M$\,}} | |
| 1074 | -\put(12,0){\line(0,1){9}\vector(0,-1){3}} | |
| 1075 | -\put(0,9){\line(1,0){12}} | |
| 1076 | -\put(0,-3){\line(1,0){12}} | |
| 1077 | -\put(0.2,4.5){\makebox(0,0)[l]{\texttt{height}}} | |
| 1078 | -\put(12.2,-1.5){\makebox(0,0)[l]{\texttt{depth}}} | |
| 1079 | -\put(6,0.2){\makebox(0,0)[b]{\texttt{width}}} | |
| 1080 | - | |
| 1081 | -\thicklines | |
| 1082 | -\put(3,0){\line(0,1){7}\line(0,-1){2.5}\line(1,0){6}} | |
| 1083 | -\put(9,0){\line(0,1){7}\line(0,-1){2.5}} | |
| 1084 | -\put(3,7){\line(1,0){6}} | |
| 1085 | -\put(3,-2.5){\line(1,0){6}} | |
| 1086 | -\newsavebox{\eqdist} | |
| 1087 | -\savebox{\eqdist}(0,0)[c]{% | |
| 1088 | - \thinlines | |
| 1089 | - \put(-0.08,0.2){\line(0,-1){0.4}}% | |
| 1090 | - \put(0.08,0.2){\line(0,-1){0.4}}} | |
| 1091 | -\put(1.5,0){\usebox{\eqdist}} | |
| 1092 | -\put(10.5,0){\usebox{\eqdist}} | |
| 1093 | - | |
| 1094 | -\thicklines | |
| 1095 | -\put(3,-1.5){\vector(-1,0){4}} | |
| 1096 | -\put(1,-1.7){\makebox(0,0)[t]{\texttt{left}}} | |
| 1097 | -\put(3,0){\vector(0,-1){1.5}} | |
| 1098 | -\put(3.2,-0.75){\makebox(0,0)[l]{\texttt{down}}} | |
| 1099 | -\end{picture} | |
| 1100 | -\end{center} | |
| 1101 | -\caption{The position of the `real' glyph.} | |
| 1102 | -\label{fig-pos} | |
| 1103 | -\end{figure} | |
| 1104 | - | |
| 1105 | -Figure~\ref{fig-pos} shows the adjustment process. A large square $M$ is | |
| 1106 | -the imaginary body specified in the metric, and a vertical | |
| 1107 | -rectangle is the imaginary body of a real glyph. First, the real glyph | |
| 1108 | -is aligned with respect to the width of $M$. In the figure, the real | |
| 1109 | -glyph is aligned `middle'; this setting is useful for the full-width | |
| 1110 | -middle dot `・'. We have other settings, `left' and `right'. | |
| 1111 | -After that, it is shifted according to the value of |left| and |down|, | |
| 1112 | -which are specified in the metric, too. The final position of the real glyph | |
| 1113 | -is shown by the gray rectangle~$R$. If the amount of shifting the baseline is | |
| 1114 | -not zero, $M$ (and hence the real glyph) is shifted by that amount. | |
| 1115 | - | |
| 1116 | -We would like to remark briefly on the vertical position of a real | |
| 1117 | -glyph. A JFM (or a metric used in \LuaTeX-ja) and a real font used for | |
| 1118 | -it may have different height or depth. In that case, it may look better | |
| 1119 | -if the real glyph is shifted vertically to match the height-depth ratio | |
| 1120 | -specified in the metric, while any vertical adjustment except the | |
| 1121 | -adjustment by the |down| value does not performed in the present | |
| 1122 | -implementation of \LuaTeX-ja . This situation is carefully studied by | |
| 1123 | -Otobe~\cite{min10}. Here the policy on this problem is not determined | |
| 1124 | -now, however we would like to offer several solutions in future | |
| 1125 | -development. | |
| 1126 | - | |
| 1127 | -\section{Conclusion} | |
| 1128 | -We have discussed about our \LuaTeX-ja package, which is much affected | |
| 1129 | -by \pTeX. For now, it can be used for experimental use, however there | |
| 1130 | -are much refinements which are needed for regular use. The author hopes | |
| 1131 | -that this paper and \LuaTeX-ja project contribute the typesetting Japanese, | |
| 1132 | -and possibly other Asian languages, under \LuaTeX. | |
| 1133 | - | |
| 1134 | -\section*{Acknowledgements} | |
| 1135 | -The author would like to thank Ken Nakano and Hideaki Togashi for their | |
| 1136 | -development of ASCII \pTeX. The author is very grateful to Haruhiko | |
| 1137 | -Okumura for his leadership in the Japanese \TeX\ community. The author | |
| 1138 | -is also very grateful to members of \LuaTeX-ja project team for their | |
| 1139 | -valuable cooperation in development. | |
| 1140 | - | |
| 1141 | -%%% The style of the bibiliogrphy is `amsplain'. | |
| 1142 | -\providecommand{\bysame}{\leavevmode\hbox to3em{\hrulefill}\thinspace} | |
| 1143 | -\providecommand{\href}[2]{#2} | |
| 1144 | -\begin{thebibliography}{99} | |
| 1145 | - | |
| 1146 | -\bibitem{aj16} | |
| 1147 | -Adobe Systems Incorporated, \emph{Adobe-Japan1-6 Character Collection | |
| 1148 | - for CID-Keyed Fonts}, Technical Note~\#5078, 2004. | |
| 1149 | -\url{http://partners.adobe.com/public/developer/en/font/5078.Adobe-Japan1-6.pdf} | |
| 1150 | - | |
| 1151 | -\bibitem{ptex} | |
| 1152 | -ASCII MEDIA WORKS，アスキー日本語\TeX\ (\pTeX)．\url{http://ascii.asciimw.jp/pb/ptex/} | |
| 1153 | - | |
| 1154 | -\bibitem{apl} | |
| 1155 | -John Baker, \emph{Typesetting UTF8 APL code with the \LaTeX\ lstlisting package}. | |
| 1156 | -\url{http://bakerjd99.wordpress.com/2011/08/15/} | |
| 1157 | - | |
| 1158 | -\bibitem{omega} | |
| 1159 | -Jin-Hwan~Cho and Haruhiko Okumura, \emph{Typesetting CJK Languages with Omega}, | |
| 1160 | -\TeX, XML, and Digital Typography, Lecture Notes in Computer Science, vol.~3130, | |
| 1161 | -Springer, 2004, 139--148. | |
| 1162 | - | |
| 1163 | -\bibitem{joylua} | |
| 1164 | -Yannis Haralambous. \emph{The Joy of \LuaTeX}. \url{http://luatex.bluwiki.com/} | |
| 1165 | - | |
| 1166 | -\bibitem{jisx4051} | |
| 1167 | -Japanese Industrial Standards Committee. \emph{JIS~X~4051: Formatting | |
| 1168 | - rules for Japanese documents}, 1993, 1995, 2004. | |
| 1169 | - | |
| 1170 | -\bibitem{eptex} | |
| 1171 | -北川弘典，$\varepsilon$-\pTeX についてのwiki． | |
| 1172 | -\url{http://sourceforge.jp/projects/eptex/wiki/FrontPage} | |
| 1173 | - | |
| 1174 | -\bibitem{luaums} | |
| 1175 | -北川弘典，\LuaTeX で日本語． | |
| 1176 | -\url{http://oku.edu.mie-u.ac.jp/tex/mod/forum/discuss.php?d=378} | |
| 1177 | - | |
| 1178 | -\bibitem{luatexref} | |
| 1179 | -\LuaTeX\ development team, \emph{The \LuaTeX\ reference}. | |
| 1180 | -\url{http://www.luatex.org/svn/trunk/manual/luatexref-t.pdf} (snapshot of SVN trunk) | |
| 1181 | - | |
| 1182 | -\bibitem{man} | |
| 1183 | -\LuaTeX-ja project team, \emph{The \LuaTeX-ja package}. | |
| 1184 | -Not completed for now. Available at |doc/man-en.pdf| (in English) or | |
| 1185 | - |doc/man-ja.pdf| (in Japanese) | |
| 1186 | -in the Git repository. | |
| 1187 | - | |
| 1188 | -\bibitem{luajp-test} | |
| 1189 | -香田温人，\LuaTeX と日本語． | |
| 1190 | -\url{http://www1.pm.tokushima-u.ac.jp/~kohda/tex/luatex-old.html} | |
| 1191 | - | |
| 1192 | -\bibitem{luajalayout} | |
| 1193 | -前田一貴，luajalayout パッケージ---Lua\LaTeX によ | |
| 1194 | - る日本語組版---． | |
| 1195 | -\url{http://www-is.amp.i.kyoto-u.ac.jp/lab/kmaeda/lualatex/luajalayout/} | |
| 1196 | - | |
| 1197 | -\bibitem{jsclasses} | |
| 1198 | -奥村晴彦，p\LaTeXe 新ドキュメントクラス． | |
| 1199 | -\url{http://oku.edu.mie-u.ac.jp/~okumura/jsclasses/} | |
| 1200 | - | |
| 1201 | -\bibitem{ptexjp} | |
| 1202 | -Haruhiko Okumura, \emph{\pTeX\ and Japanese Typesetting}, | |
| 1203 | - The Asian Journal of \TeX\ \textbf{2}~(2008), 43--51. | |
| 1204 | - | |
| 1205 | -\bibitem{min10} | |
| 1206 | -乙部厳己，min10フォントについて． | |
| 1207 | -\url{http://argent.shinshu-u.ac.jp/~otobe/tex/files/min10.pdf} | |
| 1208 | - | |
| 1209 | -\bibitem{otf} | |
| 1210 | -齋藤修三郎，Open Type Font用VF． | |
| 1211 | -\url{http://psitau.kitunebi.com/otf.html} | |
| 1212 | - | |
| 1213 | -\bibitem{stack-mail} | |
| 1214 | -Jonathan Sauer, \emph{[Dev-luatex] tex.currentgrouplevel}. | |
| 1215 | -\url{http://www.ntg.nl/pipermail/dev-luatex/2008-August/001765.html} | |
| 1216 | - | |
| 1217 | -\bibitem{uptex} | |
| 1218 | -Takuji Tanaka, \emph{u\pTeX, up\LaTeX---unicode version of \pTeX, p\LaTeX}. | |
| 1219 | -\url{http://homepage3.nifty.com/ttk/comp/tex/uptex_en.html} | |
| 1220 | - | |
| 1221 | -\bibitem{ptexenc} | |
| 1222 | -Nobuyuki Tsuchimura, \emph{Development of a Japanese \TeX\ Distribution~`ptetex3'}, | |
| 1223 | -Computer Software\ \textbf{24} (2007), no.~4, 40--50, (in Japanese). | |
| 1224 | - | |
| 1225 | -\bibitem{w3c} | |
| 1226 | -W3C Working Group, \emph{Requirements for Japanese Text Layout}. | |
| 1227 | -\url{http://www.w3.org/TR/jlreq/} | |
| 1228 | -\end{thebibliography} | |
| 1229 | - | |
| 1230 | -\end{document} | |
| 1 | +%#!lualatex ajt-devel-ltja | |
| 2 | +\documentclass{ajt} | |
| 3 | + | |
| 4 | +%%% Packages used in this paper | |
| 5 | + | |
| 6 | +%%% Font setting for \LuaTeX; this is extract from ajt.cls | |
| 7 | +\makeatletter | |
| 8 | + \if@print | |
| 9 | + \RequirePackage{fontspec,xunicode} | |
| 10 | + \RequirePackage{luatextra} | |
| 11 | + \setmainfont[Mapping=tex-text]{Palatino LT Std} | |
| 12 | + \setsansfont[Mapping=tex-text]{Optima LT Std} | |
| 13 | + \else | |
| 14 | + \RequirePackage{fontspec,luatextra} | |
| 15 | + \setmainfont[Mapping=tex-text]{TeX Gyre Pagella} % \simeq Palatino | |
| 16 | + \fi | |
| 17 | + | |
| 18 | +%%% LuaTeX-ja | |
| 19 | +\usepackage{luatexja,luatexja-fontspec} | |
| 20 | +\ltjsetparameter{jacharrange={-3,-8}} | |
| 21 | +\DeclareFontShape{JY3}{mc}{m}{n}{<-> s*[0.92489] file:ipam.ttf:jfm=ujis}{} | |
| 22 | +\DeclareFontShape{JY3}{gt}{m}{n}{<-> s*[0.92489] file:ipag.ttf:jfm=ujis}{} | |
| 23 | +% quick hack: monospaced Japanese font by \ttfamily | |
| 24 | +\DeclareKanjiFamily{JY3}{\ttdefault}{}{} | |
| 25 | +\DeclareFontShape{JY3}{\ttdefault}{m}{n}{<-> s*[0.92489] file:ipag.ttf:jfm=mono}{} | |
| 26 | + | |
| 27 | + | |
| 28 | +%%% LTXexample environment | |
| 29 | +\usepackage{showexpl,lltjlisting} | |
| 30 | +\lstset{basicstyle=\ttfamily\small, width=0.3\textwidth, basewidth=.5em} | |
| 31 | + | |
| 32 | +%%% Verbatim environment | |
| 33 | +\usepackage{fancyvrb} | |
| 34 | +\CustomVerbatimEnvironment{code}{Verbatim}% | |
| 35 | +{numbers=left,xleftmargin=1.5em,baselinestretch=1.069,fontsize=\small} | |
| 36 | +\CustomVerbatimEnvironment{codewithoutnum}{Verbatim}% | |
| 37 | +{xleftmargin=1.5em,baselinestretch=1.069,fontsize=\small} | |
| 38 | +\CustomVerbatimEnvironment{codewithoutnumsmall}{Verbatim}% | |
| 39 | +{xleftmargin=1.5em,baselinestretch=1.0,fontsize=\footnotesize} | |
| 40 | +\DefineShortVerb{\|} | |
| 41 | + | |
| 42 | +%%% Others | |
| 43 | +\usepackage{mflogo,booktabs} | |
| 44 | +\definecolor{grayx}{gray}{0.85} | |
| 45 | +\hyphenation{ | |
| 46 | + kanjiskip | |
| 47 | + xkanjiskip | |
| 48 | +} | |
| 49 | + | |
| 50 | +%%% Mandatory article metadata %%% | |
| 51 | +\title{Development of \LuaTeX-ja package} | |
| 52 | +\author{Hironori Kitagawa {\normalsize 北川 弘典}} | |
| 53 | +\address{\LuaTeX-ja project team} | |
| 54 | +\email{h\_kitagawa2001@yahoo.co.jp} | |
| 55 | + | |
| 56 | +\keywords{\TeX, p\TeX, \LuaTeX, \LuaTeX-ja, Japanese} | |
| 57 | +\abstract{% | |
| 58 | +\LuaTeX-ja package is a macro package for typesetting Japanese | |
| 59 | +documents under \LuaTeX. The package has more flexibility of | |
| 60 | +typesetting than \pTeX, which is widely used Japanese extension of \TeX, | |
| 61 | +and has corrected some unwanted features of \pTeX. | |
| 62 | +In this paper, we describe specifications, the current status and some | |
| 63 | +internal processing methods of \LuaTeX-ja. | |
| 64 | +} | |
| 65 | + | |
| 66 | +\newcommand{\parname}[1]{\textsf{#1}} | |
| 67 | +\newcommand{\jstrut}{\vrule width0pt height\cht depth\cdp} | |
| 68 | +\newcommand{\imagfm}[1]{\ifvmode\leavevmode\fi% | |
| 69 | + \hbox{\fboxsep=0pt\fbox{\setbox0=\hbox{#1}\copy0\kern-\wd0 | |
| 70 | + \smash{\vrule width \wd0 height 0.4pt depth0.4pt}}}} | |
| 71 | +\begin{document} | |
| 72 | + | |
| 73 | +%%% Do not forget to start with \maketitle! | |
| 74 | +\maketitle | |
| 75 | + | |
| 76 | +\section{Introduction} | |
| 77 | +\subsection{History} | |
| 78 | +To typeset Japanese documents with \TeX, ASCII \pTeX~\cite{ptex} has | |
| 79 | +been widely used in Japan. There are other methods---for example, using | |
| 80 | +Omega and OTP~\cite{omega}, or with the CJK package---to do so, however, | |
| 81 | +these alternative methods did not become majority. The author thinks | |
| 82 | +that this is because \pTeX\ enables us to produce high-quality documents | |
| 83 | +(e.g.,~supporting vertical typesetting), and the appearance of \pTeX\ is | |
| 84 | +earlier than that of alternatives described above. | |
| 85 | + | |
| 86 | +However, \pTeX\ has been left behind from the extensions of \TeX\ such | |
| 87 | +as \eTeX\ and \pdfTeX, and the diffusion of UTF-8 encoding. In recent | |
| 88 | +years, the situation has become better, by development of | |
| 89 | +|ptexenc|~\cite{ptexenc} by Nobuyuki Tsuchimura (\hbox{土村展之}), | |
| 90 | +$\varepsilon$-\pTeX~\cite{eptex} by the author,~and u\pTeX~\cite{uptex} | |
| 91 | +by Takuji Tanaka (田中琢爾). However, continuing this approach, namely, | |
| 92 | +to develop an engine extension localized for Japanese, is not wise. This | |
| 93 | +approach needs lots of work for \emph{each} engine. In addition, if we | |
| 94 | +use \LuaTeX, the necessity of an engine extension is getting smaller | |
| 95 | +because \LuaTeX\ has an ability to hook \TeX's internal process by using | |
| 96 | +Lua callbacks. | |
| 97 | + | |
| 98 | + | |
| 99 | +There were several experimental attempts to typeset | |
| 100 | +Japanese documents with \LuaTeX\ before. Here we cite three examples: | |
| 101 | +\begin{itemize} | |
| 102 | +\item |luaums.sty|~\cite{luaums} developed by the author. This | |
| 103 | + experimental package is for creating a certain Japanese-based presentation | |
| 104 | + with \LuaTeX. | |
| 105 | +\item the \emph{luajalayout} package~\cite{luajalayout}, formerly known as the | |
| 106 | + \emph{jafontspec} package, by Kazuki Maeda (前田一貴). This package is based on | |
| 107 | + \LaTeXe\ and \emph{fontspec} package. | |
| 108 | +\item the \emph{luajp-test} package~\cite{luajp-test}, a test package made by | |
| 109 | + Atsuhito Kohda (香田温人), based on articles on the web page~\cite{joylua}. | |
| 110 | +\end{itemize} | |
| 111 | +However, these packages are based on \LaTeXe, and do not have much | |
| 112 | +ability to control the typesetting rule. And it is inefficient that more | |
| 113 | +than one people separately develop similar packages. Development of the | |
| 114 | +\LuaTeX-ja package is started initially by the author and Kazuki Maeda, because of | |
| 115 | +these situations. | |
| 116 | + | |
| 117 | +\subsection{Development policy of \LuaTeX-ja} | |
| 118 | +\label{ssec-pol} | |
| 119 | +The first aim of \LuaTeX-ja project was to implement features (from the | |
| 120 | +`primitive' level) of \pTeX\ as macros under \LuaTeX, therefore \LuaTeX-ja is | |
| 121 | +much affected by \pTeX. However, as development proceeded, some | |
| 122 | +technical/conceptual difficulties arose. Hence we changed the aim | |
| 123 | +of the project as follows: | |
| 124 | +\begin{itemize} | |
| 125 | +\item\emph{\LuaTeX-ja offers at least the same flexibility of | |
| 126 | + typesetting that p\TeX\ has.} | |
| 127 | + | |
| 128 | + We are not satisfied with the ability of producing outputs conformed to | |
| 129 | + JIS~X~4051~\cite{jisx4051}, the Japanese Industrial Standard for | |
| 130 | + typesetting, or to a technical note~\cite{w3c} by W3C; | |
| 131 | + if one wants to produce very incoherent outputs for some reason, it | |
| 132 | + should be possible. | |
| 133 | +In this point, previous attempts of Japanese typesetting with \LuaTeX\ | |
| 134 | + which we cited in the previous subsection are inadequate. | |
| 135 | + | |
| 136 | +\pTeX\ has some flexibility of typesetting, by changing internal | |
| 137 | + parameters such as |\kanjiskip| or |\prebreakpenalty|, and by using | |
| 138 | + custom JFM (Japanese TFM). Therefore we decided to include these | |
| 139 | + functionality to \LuaTeX-ja. | |
| 140 | + | |
| 141 | +\item\emph{\LuaTeX-ja isn't mere re-implementation or porting of \pTeX; | |
| 142 | + some (technically and/or conceptually) inconvenient features of | |
| 143 | + \pTeX\ are modified.} | |
| 144 | + | |
| 145 | + We describe this point in more detail at the next section. | |
| 146 | +\end{itemize} | |
| 147 | + | |
| 148 | + | |
| 149 | +\subsection{Overview of the processes} | |
| 150 | +\label{ssec-over} | |
| 151 | +We describe an outline of \LuaTeX-ja's process in order. | |
| 152 | + | |
| 153 | +\begin{itemize} | |
| 154 | +\item In the |process_input_buffer| callback: treatment of breaking | |
| 155 | + lines after a Japanese character (in Subsection~\ref{ssec-line}). | |
| 156 | + | |
| 157 | +\item In the |hyphenate| callback: font replacement. | |
| 158 | + | |
| 159 | +\LuaTeX-ja looks into for each \textit{glyph\_node}~$p$ in the horizontal list. If | |
| 160 | + the character represented by $p$ is considered as a Japanese | |
| 161 | + character, the font used at $p$ is replaced by the value of | |
| 162 | + |\ltj@curjfnt|, an attribute for `the current Japanese font' | |
| 163 | + at~$p$. | |
| 164 | + | |
| 165 | +Furthermore, the subtype of $p$ is subtracted by 1 to suppress | |
| 166 | + hyphenation around $p$ by \LuaTeX, because later processes of | |
| 167 | + \LuaTeX-ja take care of all things about Japanese characters. | |
| 168 | + | |
| 169 | +\item In |pre_linebreak_filter| and |hpack_filter| callbacks: | |
| 170 | + | |
| 171 | +\begin{enumerate} | |
| 172 | +\item \LuaTeX-ja has its own stack system, and the current horizontal | |
| 173 | + list is traversed in this stage to determine what the level of | |
| 174 | + \LuaTeX-ja's internal stack at the end of the list is. We will | |
| 175 | + discuss it in Subsection~\ref{ssec-stack}. | |
| 176 | + | |
| 177 | +\item In this stage, \LuaTeX-ja inserts glues/kerns for Japanese | |
| 178 | + typesetting in the list. This is the core routine of \LuaTeX-ja. | |
| 179 | + We will discuss it in Subsections | |
| 180 | + \ref{ssec-jglue}~and~\ref{ssec-jspec} . | |
| 181 | + | |
| 182 | +\item To make a match between a metric and a real font, sometimes | |
| 183 | + adjustument of the position of (Japanese) glyphs are performed. | |
| 184 | + We will discuss it in Subsection~\ref{ssec-width}. | |
| 185 | +\end{enumerate} | |
| 186 | +\item In the |mlist_to_hlist| callback: treatment of Japanese characters | |
| 187 | + in math formulas. This stage is similar to adjustment of the | |
| 188 | + position of glyphs (see above), so we omit to describe this stage | |
| 189 | + from this paper. | |
| 190 | +\end{itemize} | |
| 191 | + | |
| 192 | +In this paper, a \emph{alphabetic character} means a non-Japanese | |
| 193 | +character. Similarly, we use the word an \emph{alphabetic font} as the | |
| 194 | +counterpart of a jJpanese font. | |
| 195 | + | |
| 196 | +\subsection{Contents of this paper} | |
| 197 | +Here we describe the contents of the rest of this paper briefly. In | |
| 198 | +Section~\ref{sec:differences_with_ptex}, we describe major differences | |
| 199 | +between \pTeX\ and \LuaTeX-ja. The next section, | |
| 200 | +Section~\ref{sec:distinction_of_characters}, is concentrated on a | |
| 201 | +problem how we distinguish between Japanese characters and alphabetic | |
| 202 | +characters. In Section~\ref{sec:current_status}, we show current | |
| 203 | +development status of the package. Finally, in | |
| 204 | +Section~\ref{sec:implementation}, we describe some internal routines of | |
| 205 | +\LuaTeX-ja. | |
| 206 | + | |
| 207 | +\subsection{General information of the project} | |
| 208 | +This \LuaTeX-ja project is hosted by SourceForge.jp. The official wiki | |
| 209 | +is located on | |
| 210 | +\url{http://sourceforge.jp/projects/luatex-ja/wiki/}. There is | |
| 211 | +no stable version on October 22, 2011, however a set of developer sources can be | |
| 212 | +obtained from the git repository. Members of the project team are as follows | |
| 213 | +(in random order): Hironori Kitagawa, Kazuki Maeda, Takayuki Yato, | |
| 214 | +Yusuke Kuroki, Noriyuki Abe, Munehiro Yamamoto, Tomoaki Honda, | |
| 215 | +and~Shuzaburo Saito. | |
| 216 | + | |
| 217 | + | |
| 218 | +\section{Major differences with \pTeX} | |
| 219 | +\label{sec:differences_with_ptex} | |
| 220 | +In this section, we explain several major differences between \pTeX\ | |
| 221 | +and our \LuaTeX-ja. For general information of Japanese typesetting and the | |
| 222 | +overview of \pTeX, please see Okumura~\cite{ptexjp}. | |
| 223 | + | |
| 224 | + | |
| 225 | +\subsection{Names of control sequences} | |
| 226 | +\label{ssec-csname} Because \pTeX\ is an engine modification of Knuth's | |
| 227 | +original \TeX82 engine, some of the additional primitives take a form that is | |
| 228 | +very difficult to be simulated by a macro. For example, an additional | |
| 229 | +primitive |\prebreakpenalty|$\langle\hbox{\it | |
| 230 | +char\_code}\rangle$|[=]|$\langle\hbox{\it penalty}\rangle$ in \pTeX\ | |
| 231 | +sets the amount of penalty inserted before a character whose code is | |
| 232 | +$\langle\hbox{\it char\_code}\rangle$ to $\langle\hbox{\it | |
| 233 | +penalty}\rangle$, and this form |\prebreakpenalty|$\langle\hbox{\it | |
| 234 | +char\_code}\rangle$ can be also used for retrieving the value. | |
| 235 | + | |
| 236 | +Moreover, there are some internal parameters of \pTeX\ which values of them at the end of a | |
| 237 | +horizontal box or that of a paragraph are valid in whole box or | |
| 238 | +paragraph. However, the implementation of these parameters in | |
| 239 | +\LuaTeX-ja is not so easy; we will discuss it in Subsection~\ref{ssec-stack}. | |
| 240 | + | |
| 241 | +From above two problems discussed above, the assignment and retrieval | |
| 242 | +of most parameters in \LuaTeX-ja are summarized into the following | |
| 243 | +three control sequences: | |
| 244 | +\begin{itemize} | |
| 245 | +\item |\ltjsetparameter{|$\langle\hbox{\it | |
| 246 | + name}\rangle$|=|$\langle\hbox{\it value}\rangle$|,...}|: for local | |
| 247 | + assignment. | |
| 248 | +\item |\ltjglobalsetparameter|: for global assignment. Note that these two control | |
| 249 | + sequences obey the value of |\globaldefs| primitive. | |
| 250 | +\item |\ltjgetparameter{|$\langle\hbox{\it | |
| 251 | + name}\rangle$|}[{|$\langle\hbox{\it optional | |
| 252 | + argument}\rangle$|}]|: for retrieval. The returned value is always | |
| 253 | + a string. | |
| 254 | +\end{itemize} | |
| 255 | + | |
| 256 | +\subsection{Line-break after a Japanese character} | |
| 257 | +\label{ssec-line} | |
| 258 | + | |
| 259 | +Japanese texts can break lines almost everywhere, in contrast with | |
| 260 | +alphabetic texts can break lines only between words (or use | |
| 261 | +hyphenation). Hence, \pTeX's input processor is modified so that a | |
| 262 | +line-break after a Japanese character doesn't emit a space. However, | |
| 263 | +there is no way to customize the input processor of \LuaTeX, other than | |
| 264 | +to hack its CWEB-source. All a macro package can do is to modify an input line before | |
| 265 | +when \LuaTeX\ begin to process it, inside the |process_input_buffer| | |
| 266 | +callback. | |
| 267 | + | |
| 268 | +Hence, in \LuaTeX-ja, a comment letter (we reserve U+FFFFF for this | |
| 269 | +purpose) will be appended to an input line, if this line ends with a Japanese | |
| 270 | +character.\footnote{Strictly speaking, it also requires that the catcode | |
| 271 | +of the end-line character is 5~(\emph{end-of-line}). This condition is | |
| 272 | +useful under the verbatim environment.} One might jump to a conclusion | |
| 273 | +that the treatment of a line-break by \pTeX\ and that of \LuaTeX-ja are | |
| 274 | +totally same, however they are different in the respect that \LuaTeX-ja's | |
| 275 | +judgement whether a comment letter will be appended the line is done | |
| 276 | +\emph{before} the line is actually processed by \LuaTeX. | |
| 277 | + | |
| 278 | +Figure~\ref{fig-linebreak} shows an example of this situation; the | |
| 279 | +command at the first line marks most of Japanese characters as | |
| 280 | +`non-Japanese characters'. In other words, from that command onward, the | |
| 281 | +letter `あ' will be treated as an alphabetic character by | |
| 282 | +\LuaTeX-ja. Then, it is natural to have a space between `あ' and `y' in | |
| 283 | +the output, where the actual output in the figure does not so. This is | |
| 284 | +because `あ' is considered a Japanese character by \LuaTeX-ja, | |
| 285 | +when \LuaTeX-ja does the decision whether U+FFFFF will be added to the | |
| 286 | +input line~2. | |
| 287 | + | |
| 288 | +\begin{figure} | |
| 289 | +\begin{LTXexample} | |
| 290 | +\font\x=IPAMincho \x | |
| 291 | +\ltjsetparameter{jacharrange={-6}}xあ | |
| 292 | +y | |
| 293 | +\end{LTXexample} | |
| 294 | +\caption{A notable sample showing the treatment of a line-break after a | |
| 295 | +Japanese character.}\label{fig-linebreak} | |
| 296 | +\end{figure} | |
| 297 | + | |
| 298 | +\subsection{Separation between `real' fonts and metrics} | |
| 299 | +\label{ssec-sepmet} | |
| 300 | + | |
| 301 | +Traditionally, most Japanese fonts used in typesetting are not | |
| 302 | +proportional, that is, most glyphs have same size (in most cases, | |
| 303 | +square-shaped). Hence, it is not rare that the contents of different | |
| 304 | +JFMs are essentially same, and only differ in their names. For example, | |
| 305 | +|min10.tfm| and |goth10.tfm|, which are JFMs shipped with \pTeX\ for | |
| 306 | +seriffed \emph{mincho} family and sans-seriffed \emph{gothic} family, | |
| 307 | +differ their |FAMILY| and |FACE| only. Moreover, |jis.tfm| and | |
| 308 | +|jisg.tfm|, which is included in the \emph{jis} font metric, which is | |
| 309 | +used in \emph{jsclasses}~\cite{jsclasses} by Haruhiko Okumura (奥村晴彦), | |
| 310 | +are totally same as binary files. Considering this situation, we | |
| 311 | +decided to separate `real' fonts and metrics used for them in | |
| 312 | +\LuaTeX-ja. Typical declarations of Japanese fonts in the style of plain | |
| 313 | +\TeX\ are shown in Figure~\ref{fig-jfdef}. We would like to add several | |
| 314 | +remarks: | |
| 315 | +\begin{itemize} | |
| 316 | +\item A control sequence |\jfont| must be used for Japanese fonts, instead of |\font|. | |
| 317 | +\item \LuaTeX-ja automatically loads the \emph{luaotfload} package, so | |
| 318 | + \hbox{\tt file:} and \hbox{\tt name:} prefixes, and various font features can be | |
| 319 | + used as the first line in Figure~\ref{fig-jfdef}. | |
| 320 | +\item The |jfm| key specifies the metric for the font. In | |
| 321 | + Figure~\ref{fig-jfdef}, both fonts will use a metric stored in a | |
| 322 | + Lua script named |jfm-ujis.lua|. This metric is the standard | |
| 323 | + metric in \LuaTeX-ja, and is based on JFMs used in the \emph{otf} | |
| 324 | + package~\cite{otf}. | |
| 325 | +\item The \hbox{psft:} prefix can be used to specify name-only, non-embedded | |
| 326 | + fonts. When one displays a pdf with these fonts, actual fonts which | |
| 327 | + will be used for them depend on a pdf reader. | |
| 328 | +\end{itemize} | |
| 329 | +The specification of a metric for \LuaTeX-ja is similar to that of a JFM | |
| 330 | +(see \cite{ptexjp}); characters are grouped into several classes, the | |
| 331 | +size information of characters are specified for each class, and | |
| 332 | +glue/kern insertions are specified for each pair of classes. Although | |
| 333 | +the author have not tried, it may be possible to develop a program that | |
| 334 | +`converts' a JFM to a metric for \LuaTeX-ja. \LuaTeX-ja offers three | |
| 335 | +metrics by default; |jfm-ujis.lua|, |jfm-jis.lua| based on the | |
| 336 | +\emph{jis} font metric, and |jfm-min.lua| based on old |min10.tfm|. | |
| 337 | + | |
| 338 | + Note that |-kern| in features | |
| 339 | +is important, because kerning information from a real font itself will | |
| 340 | +clash with glue/kern information from the metric. | |
| 341 | + | |
| 342 | +\begin{figure} | |
| 343 | +\begin{verbatim} | |
| 344 | +\jfont\foo=file:ipam.ttf:jfm=ujis;script=latn;-kern;+jp04 at 12pt | |
| 345 | +\jfont\bar=psft:Ryumin-Light:jfm=ujis at 10pt | |
| 346 | +\end{verbatim} | |
| 347 | +\caption{Typical declarations of Japanese fonts.} | |
| 348 | +\label{fig-jfdef} | |
| 349 | +\end{figure} | |
| 350 | + | |
| 351 | +\subsection{Insertion of glues/kerns for Japanese typesetting: timing} | |
| 352 | +\label{ssec-jglue} | |
| 353 | + | |
| 354 | +As described in \cite{luatexref}, \LuaTeX's kerning and ligaturing | |
| 355 | +processes are totally different from those of \TeX82. \TeX82's process is | |
| 356 | +done just when a (sequence of) character is appended to the current | |
| 357 | +list. Thus we can interrupt this process by writing as | |
| 358 | +|f{}irm|. However, \LuaTeX's process is \emph{node-based}, that is, the | |
| 359 | +process will be done when a horizontal box or a paragraph is ended, so | |
| 360 | +|f{}irm| and |firm| yield same outputs under \LuaTeX. | |
| 361 | + | |
| 362 | +The situation for Japanese characters is more complicated. | |
| 363 | +Glues (and kerns) which are needed for Japanese | |
| 364 | +typesetting are divided into the following three categories: | |
| 365 | +\begin{itemize} | |
| 366 | +\item Glue (or kern) from the metric of Japanese fonts (\emph{JFM glue}, | |
| 367 | + for short). | |
| 368 | + | |
| 369 | +\item Default glue between a Japanese character and an alphabetic | |
| 370 | + character (\emph{xkanjiskip}, for short), usually 1/4 of | |
| 371 | + full-width (\emph{shibuaki}) with some stretch and shrink for | |
| 372 | + justifying each line. | |
| 373 | +\item Default glue between two consecutive Japanese characters | |
| 374 | + (\emph{kanjiskip}, for short). The main reason of this glue is to | |
| 375 | + enable breaking lines almost everywhere in Japanese texts. In most | |
| 376 | + cases, its natural width is zero, and some stretch/shrink for | |
| 377 | + justifying each line. | |
| 378 | +\end{itemize} | |
| 379 | +In \pTeX, these three kinds of glues are treated differently. A JFM glue | |
| 380 | +is inserted when a (sequence of) Japanese character is appended to the | |
| 381 | +current list, same as the case of alphabetic characters in \TeX82. This | |
| 382 | +means that one can interrupt the insertion process by saying |{}|. A | |
| 383 | +\emph{xkanjiskip} is inserted just before `hpack' or line-breaking of a | |
| 384 | +paragraph; this timing is somewhat similar to that of \LuaTeX's kerning | |
| 385 | +process. Finally, A \emph{kanjiskip} is not appeared as a node anywhere; | |
| 386 | +only appears implicitly in calculation of the width of a horizontal box, | |
| 387 | +that of breaking lines, and the actual output process to a DVI | |
| 388 | +file. These specifications have made \pTeX's behavior very hard to | |
| 389 | +understand. | |
| 390 | + | |
| 391 | +\LuaTeX-ja inserts glues in all three categories simultaneously inside | |
| 392 | +|hpack_filter| and |pre_linebreak_filter| callbacks. The reasons of | |
| 393 | +this specification are to behave like alphabetic characters in \LuaTeX\ | |
| 394 | +(as described in the first paragraph in this subsection), and to clarify the specification | |
| 395 | +for \LuaTeX-ja's process. | |
| 396 | + | |
| 397 | +\subsection{Insertion of glues/kerns for Japanese typesetting: specification} | |
| 398 | +\label{ssec-jspec} | |
| 399 | + | |
| 400 | +\begin{table} | |
| 401 | +\caption{Examples of differences between \pTeX\ and \LuaTeX-ja.} | |
| 402 | +\label{tab-jfmglue} | |
| 403 | +\begin{center} | |
| 404 | +\begin{tabular}{llllllll} | |
| 405 | +\toprule | |
| 406 | +&\multicolumn{1}{c}{(1)}&\multicolumn{1}{c}{(2)}&\multicolumn{1}{c}{(3)}&\multicolumn{1}{c}{(4)}\\ | |
| 407 | +Input &|あ】{}【〙\/〘| &|い』\/a| &|う）\hbox{}（| &|え］\special{}［|\\\midrule | |
| 408 | +\pTeX &あ】\hbox{}【〙\hbox{}〘&い』\/a &う）\hbox{}（ &え］\hbox{}［\\ | |
| 409 | +\LuaTeX-ja &あ】{}【〙\/〘 &い』\/a &う）\hbox{}（ &え］\special{}［\\ | |
| 410 | +\bottomrule | |
| 411 | +\end{tabular} | |
| 412 | +\end{center} | |
| 413 | +\end{table} | |
| 414 | + | |
| 415 | +\begin{figure} | |
| 416 | +\begin{center} | |
| 417 | +\fontsize{40}{40}\selectfont | |
| 418 | +\imagfm{\jstrut あ}% | |
| 419 | +\imagfm{\jstrut 】\inhibitglue}% | |
| 420 | +\imagfm{\jstrut\kern.5\zw}% | |
| 421 | +\imagfm{\jstrut\kern.5\zw}% | |
| 422 | +\imagfm{\jstrut\inhibitglue【}% | |
| 423 | +\imagfm{\jstrut 〙\inhibitglue}% | |
| 424 | +\imagfm{\jstrut\kern.5\zw}% | |
| 425 | +\imagfm{\jstrut\kern.5\zw}% | |
| 426 | +\imagfm{\jstrut\inhibitglue〘}% | |
| 427 | +\end{center} | |
| 428 | +\caption{Detail of (1) in Table~\ref{tab-jfmglue}.} | |
| 429 | +% \caption{Details of the output of \pTeX in the case of (1) in Table~\ref{tab-jfmglue}.} のほうがよい？ | |
| 430 | +\label{fig-ptexjfm} | |
| 431 | +\end{figure} | |
| 432 | + | |
| 433 | +Now we will take a look at the insertion process itself through four points. | |
| 434 | + | |
| 435 | +\begin{description} | |
| 436 | +\item[Ignored Nodes] | |
| 437 | +As noted in the previous subsection, the insertion process in \pTeX\ can | |
| 438 | + be interrupted by saying |{}| or anything else.\footnote{This | |
| 439 | + is why some tricks like \texttt{ちょ\char`\{\char`\}っと} for | |
| 440 | + \texttt{min10.tfm} and other `old' JFMs work.} This leads the | |
| 441 | + second row in Table~\ref{tab-jfmglue}, or | |
| 442 | + Figure~\ref{fig-ptexjfm}. Here `the process is interrupted' | |
| 443 | + means that \pTeX\ does not think the letter `】\inhibitglue' | |
| 444 | + is followed by `\inhibitglue【', hence two half-width glues | |
| 445 | + are inserted between `】\inhibitglue' and `\inhibitglue【', | |
| 446 | + where the left one is from `】\inhibitglue' and the right one | |
| 447 | + is from `\inhibitglue【'. | |
| 448 | + | |
| 449 | + On the other hand, in \LuaTeX-ja, the process is done inside | |
| 450 | + |hpack_filter| and |pre_linebreak_filter| callbacks. Hence, | |
| 451 | + \emph{anything that does not make any node will be | |
| 452 | + ignored}\ in \LuaTeX-ja, as shown in (1) in | |
| 453 | + Table~\ref{tab-jfmglue}. \LuaTeX-ja also ignores any nodes | |
| 454 | + which does not make any contribution to current horizontal | |
| 455 | + list---\emph{ins\_node}, \emph{adjust\_node}, | |
| 456 | + \emph{mark\_node}, \emph{whatsit\_node} and | |
| 457 | + \emph{penalty\_node}---, as shown in (4). | |
| 458 | + | |
| 459 | + | |
| 460 | +By the way, around a \emph{glyph\_node} $p$ there may be some nodes | |
| 461 | + attached to~$p$. These are an accent and kerns for | |
| 462 | + moving it to the right place, and a kern from the italic | |
| 463 | + correction\footnote{\TeX82 (and \LuaTeX) does not distinguish | |
| 464 | + between explicit kern and a kern for italic correction. To | |
| 465 | + distinguish them, an additional subtype for a kern is introduced | |
| 466 | + in \pTeX. On the other hand, \LuaTeX-ja uses an additional attribute and | |
| 467 | + redefines \texttt{\char`\\/} to set this attribute.} for $p$. It is natural that | |
| 468 | + these attachments should be ignored inside the process. Hence | |
| 469 | + \LuaTeX-ja takes this approach, as the latest version of | |
| 470 | + \pTeX\ (p3.2). This explains (2) in the figure. | |
| 471 | + % - p3.2 というのは pTeX のバージョン？ (version~p3.2) くらいに補ったほうがよいかも． | |
| 472 | + % - in the figure は指している場所がわからずつらいだろう．面倒でも in Table~\reftab-jfmglue{} のほうがふさわしいだろう． | |
| 473 | + | |
| 474 | +Summerizing above, one should put an empty horizontal box |\hbox{}| to | |
| 475 | + where he/she wants to interrupt the insertion process in | |
| 476 | + \LuaTeX-ja as (3) in the figure. | |
| 477 | + % - in the figure は指している場所がわからずつらいだろう．面倒でも in Table~\reftab-jfmglue{} のほうがふさわしいだろう． | |
| 478 | + | |
| 479 | +\item[Fonts with the Same Metric] | |
| 480 | +Recall that \LuaTeX-ja separates `real' fonts and metrics, as in Subsection~\ref{ssec-sepmet}. | |
| 481 | +Consider the following input, where all Japanese fonts use same metric | |
| 482 | + (in \LuaTeX-ja), and |\gt| selects \emph{gothic} family for | |
| 483 | + the current Japanese font family: | |
| 484 | +\begin{quote} | |
| 485 | +\begin{verbatim} | |
| 486 | +明朝）\gt （ゴシック | |
| 487 | +\end{verbatim} | |
| 488 | +\end{quote} | |
| 489 | +If the above input is processed by \pTeX, because the insertion process is | |
| 490 | + interrupt by |\gt|, the result looks like | |
| 491 | +\begin{quote} | |
| 492 | +\mc 明朝）\hbox{}\gt （ゴシック | |
| 493 | +\end{quote} | |
| 494 | +However this seems to be unnatural, since two Japanese fonts in the | |
| 495 | + output use the same metric, i.e.,~the same | |
| 496 | + typesetting rule. Hence, we decided that Japanese fonts with | |
| 497 | + the same metric are treated as one font in the insertion | |
| 498 | + process of \LuaTeX-ja. Thus, the output from the above input | |
| 499 | + in \LuaTeX-ja looks like: | |
| 500 | +\begin{quote} | |
| 501 | +\mc 明朝）\gt （ゴシック | |
| 502 | +\end{quote} | |
| 503 | +One might have the situation that this default behavior is not | |
| 504 | + suitable. \LuaTeX-ja offers a way to handle this situation, but | |
| 505 | + we leave it to the manual~\cite{man}. | |
| 506 | + | |
| 507 | +\item[Fonts with Different Metrics] | |
| 508 | +The case where two consecutive Japanese characters use different metrics and/or | |
| 509 | + different size is similar. Consider the following input where | |
| 510 | + the \emph{mincho} family and the \emph{gothic} family use | |
| 511 | + different metrics: | |
| 512 | +\begin{quote} | |
| 513 | +\begin{verbatim} | |
| 514 | +漢）\gt （漢）\large （大 | |
| 515 | +\end{verbatim} | |
| 516 | +\end{quote} | |
| 517 | +As the previous paragraph, this input yields the following, by \pTeX: | |
| 518 | +\begin{quote} | |
| 519 | +\mc 漢）\hbox{}\gt （漢）\hbox{}\large （大 | |
| 520 | +\end{quote} | |
| 521 | +We had thought that amounts of spaces between parentheses in above output | |
| 522 | + are too much. Hence we have changed the default behavior of | |
| 523 | + \LuaTeX-ja, so that the amount of a glue between two Japanese | |
| 524 | + characters with different metrics is the \emph{average} of a glue | |
| 525 | + from the left character and that from the right | |
| 526 | + character. For example, Figure~\ref{fig-diffmet} shows the | |
| 527 | + output from above input. The width of glue indicated `(1)' is | |
| 528 | + $(a/2 + a/2)/2 = 0.5a$, and the width of glue indicated `(2)' | |
| 529 | + is $(a/2 + 1.2a/2)/2 = 0.55a$. This default behavior can be | |
| 530 | + changed by \textsf{diffrentmet} parameter of \LuaTeX-ja. | |
| 531 | + | |
| 532 | +\begin{figure} | |
| 533 | +\begin{center} | |
| 534 | +\fontsize{40}{40}\selectfont | |
| 535 | +\imagfm{\jstrut\smash{% | |
| 536 | + \vtop{\lineskiplimit=\maxdimen\lineskip2pt\halign{#\cr漢\cr | |
| 537 | + \small\vrule height .5ex depth .5ex\hrulefill\ \lower.5ex\hbox{$a$}\ | |
| 538 | + \hrulefill\vrule height .5ex depth .5ex\cr}}}}% | |
| 539 | +\imagfm{\jstrut ）\inhibitglue}% | |
| 540 | +\hbox to .5\zw{\hss\normalsize (1)\hss}% | |
| 541 | +\imagfm{\jstrut\inhibitglue\gt （}% | |
| 542 | +\imagfm{\jstrut\gt 漢}% | |
| 543 | +\imagfm{\jstrut\gt ）\inhibitglue}% | |
| 544 | +\hbox to .55\zw{\hss\normalsize (2)\hss}% | |
| 545 | +\imagfm{\fontsize{48}{48}\selectfont\jstrut\gt\inhibitglue （}% | |
| 546 | +\imagfm{\fontsize{48}{48}\selectfont\jstrut\smash{% | |
| 547 | + \vtop{\lineskiplimit=\maxdimen\lineskip2pt\halign{#\cr\gt 大\cr | |
| 548 | + \small\vrule height .5ex depth .5ex\hrulefill\ \lower.5ex\hbox{$1.2a$}\ | |
| 549 | + \hrulefill\vrule height .5ex depth .5ex\cr}}}} | |
| 550 | +\end{center} | |
| 551 | +\caption{Fonts with different metrics.} | |
| 552 | +\label{fig-diffmet} | |
| 553 | +\end{figure} | |
| 554 | + | |
| 555 | +\item[\emph{kanjiskip} and \emph{xkanjiskip}] | |
| 556 | +In \pTeX, the value of \emph{xkanjiskip} is controlled by a skip named | |
| 557 | + |\xkanjiskip|. A well-known defect of this implementation is | |
| 558 | + that the value of \emph{xkanjiskip} is not connected with the | |
| 559 | + size of the currnt Japanese font. It seems that |EXTRASPACE|, | |
| 560 | + |EXTRASTRETCH|, |EXTRASHRINK| parameters in a JFM are | |
| 561 | + reserved for specifying the default value of | |
| 562 | + \emph{xkanjiskip} in a unit of the design size, but \pTeX\ | |
| 563 | + did not use these parameters, actually. | |
| 564 | + | |
| 565 | +Considering this situation of p\TeX, \LuaTeX-ja can use the value of | |
| 566 | + \emph{xkanjiskip} that specified in a metric. If the value of | |
| 567 | + \emph{xkanjiskip} on user side (this is the value of | |
| 568 | + \textsf{xkanjiskip} parameter of |\ltjsetparameter|) is | |
| 569 | + |\maxdimen|, then \LuaTeX-ja use the specification from | |
| 570 | + the current used metric as the actual value of | |
| 571 | + \emph{xkanjiskip}. This description also applies for \emph{kanjiskip}. | |
| 572 | +\end{description} | |
| 573 | + | |
| 574 | +\section{Distinction of characters} | |
| 575 | +\label{sec:distinction_of_characters} Since \LuaTeX\ can handle Unicode | |
| 576 | +characters natively, it is a major problem that how we distinguish | |
| 577 | +Japanese characters and alphabetic characters. For example, the | |
| 578 | +multiplication sign (U+00D7) exists both in ISO-8859-1 (hence in Latin-1 | |
| 579 | +Supplement in Unicode) and in the basic Japanese character set | |
| 580 | +JIS~X~0208. It is not desirable that this character is always treated as | |
| 581 | +an alphabetic character, because this symbol is often used in the sense | |
| 582 | +of `negative' in Japan. | |
| 583 | + | |
| 584 | +\subsection{Character ranges} | |
| 585 | +Before we describe the approach taken is \LuaTeX-ja, we review the | |
| 586 | +approach taken by u\pTeX. u\pTeX\ extends the |\kcatcode| primitive in | |
| 587 | +\pTeX, to use this primitive for setting how a character is treated | |
| 588 | +among alphabetic characters~(15), \emph{kanji}~(16), \emph{kana}~(17), | |
| 589 | +\emph{kanji}, \emph{Hangul}~(17), or~\emph{other CJK characters}~(18). | |
| 590 | +The assignment to |\kcatcode| can be done by a Unicode | |
| 591 | +block.\footnote{There are some exceptions. For example, U+FF00--FFEF | |
| 592 | +(Halfwidth and Fullwidth Forms) are divided into three blocks in recent | |
| 593 | +u\pTeX.} | |
| 594 | + | |
| 595 | +\LuaTeX-ja adopted a different approach. There are many Unicode blocks | |
| 596 | + in Basic Multilingual Plane which are not included in | |
| 597 | + Japanese fonts, therefore it is inconvenient if we process by a Unicode | |
| 598 | + block. Furthermore, JIS~X~0208 are not just union of Unicode | |
| 599 | + blocks; for example, the intersection of JIS~X~0208 and | |
| 600 | + Latin-1 Supplement is shown in | |
| 601 | + Table~\ref{tab-inter}. Considering these two points, to | |
| 602 | + customize the range of Japanese characters in \LuaTeX-ja, one | |
| 603 | + has to define ranges of character codes in his source in advance. | |
| 604 | + | |
| 605 | + | |
| 606 | +\begin{table} | |
| 607 | +\caption{Intersection of JIS~X~0208 and Latin-1 Supplement.} | |
| 608 | +\label{tab-inter} | |
| 609 | +\begin{center} | |
| 610 | +\begin{tabular}{llll} | |
| 611 | +\ltjjachar"A7 (U+00A7),& | |
| 612 | +\ltjjachar"A8 (U+00A8),& | |
| 613 | +\ltjjachar"B0 (U+00B0),& | |
| 614 | +\ltjjachar"B1 (U+00B1),\\ | |
| 615 | +\ltjjachar"B4 (U+00B4),& | |
| 616 | +\ltjjachar"B6 (U+00B6),& | |
| 617 | +\ltjjachar"D7 (U+00D7),& | |
| 618 | +\ltjjachar"F7 (U+00F7) | |
| 619 | +\end{tabular} | |
| 620 | +\end{center} | |
| 621 | +\end{table} | |
| 622 | + | |
| 623 | +%%Example... | |
| 624 | + | |
| 625 | +We note that \LuaTeX-ja offers two additional control sequences, | |
| 626 | + |\ltjjachar| and |\ltjalchar|. They are similar to |\char| | |
| 627 | + primitive, however |\ltjjachar| always yields a Japanese character, provided that | |
| 628 | + the argument is more than or equal to 128, and |\ltjalchar| always | |
| 629 | + yields an alphabetic character, regardless of the argument. | |
| 630 | + | |
| 631 | +\subsection{Default setting of ranges} | |
| 632 | +Patches for plain \TeX\ and \LaTeXe\ of \LuaTeX-ja predefine 8~character | |
| 633 | +ranges, as shown in Table~\ref{tab-chrrng}. Almost of these ranges are | |
| 634 | +just the union of Unicode blocks, and determined from the Adobe-Japan1-6 | |
| 635 | +character collection~\cite{aj16}, and JIS~X~0208. Among these 8~ranges, | |
| 636 | +the ranges~2, 3, 6, 7, and~8 are considered ranges of Japanese | |
| 637 | +characters, and others are considered ranges of alphabetic | |
| 638 | +characters.\footnote{Note that ranges 3~and~8 are considered ranges of | |
| 639 | +alphabetic characters in this paper.} We remark on ranges 2~and~8: | |
| 640 | +\begin{description} | |
| 641 | +\item[The range~2] | |
| 642 | +JIS~X~0208 includes Greek letters and Cyrillic letters, however, these | |
| 643 | + letters cannot be used for typesetting Greek or Russian, of | |
| 644 | + course. Hence it is reasonable that Greek letters and | |
| 645 | + Cyrillic consist another character range. | |
| 646 | +\item[The range~8] | |
| 647 | +If one want to use 8-bit TFMs, such as T1 or TS1 encodings, he should | |
| 648 | + mark this range~8 as a range of alphabetic characters by | |
| 649 | +\begin{quote} | |
| 650 | +|\ltjsetparameter{jacharrange={-8}}| | |
| 651 | +\end{quote} | |
| 652 | +This is because some 8-bit TFMs have a glyph in this range; for example, | |
| 653 | + the character `\OE' is located at |"D7| in the T1 encoding. %" | |
| 654 | +\end{description} | |
| 655 | + | |
| 656 | + | |
| 657 | +\begin{table} | |
| 658 | +\caption{Predefined ranges in \LuaTeX-ja.} | |
| 659 | +\label{tab-chrrng} | |
| 660 | +\begin{center} | |
| 661 | +\begin{tabular}{@{\bf}rl} | |
| 662 | +1&(Additional) Latin characters which are not belonged in the range~8.\\ | |
| 663 | +2&Greek and Cyrillic letters.\\ | |
| 664 | +3&Punctuations and miscellaneous symbols.\\ | |
| 665 | +4&Unicode blocks which does not intersect with Adobe-Japan1-6.\\ | |
| 666 | +5&Surrogates and supplementary private use Areas.\\ | |
| 667 | +6&Characters used in Japanese typesetting.\\ | |
| 668 | +7&Characters possibly used in CJK typesetting, but not in Japanese.\\ | |
| 669 | +8&Characters in Table~\ref{tab-inter}. | |
| 670 | +\end{tabular} | |
| 671 | +\end{center} | |
| 672 | +\end{table} | |
| 673 | + | |
| 674 | +\subsection{Control sequences producing Unicode characters} | |
| 675 | +\label{ssec-unichar} | |
| 676 | + | |
| 677 | +The \emph{fontspec} package\footnote{Preciously saying, it is the | |
| 678 | +\emph{xunicode} package, originally a package for \XeTeX and | |
| 679 | +automatically loaded by the \emph{fontspec} package.} offers various | |
| 680 | +control sequences that produce Unicode characters. However, these | |
| 681 | +control sequences as it stands cannot work correctly with the default | |
| 682 | +range setting of \LuaTeX-ja. For example, |\textquotedblleft| is just | |
| 683 | +an abbreviation of |\char"201C\relax|, and the character U+201C (LEFT %" | |
| 684 | +DOUBLE QUOTATION MARK) is treated as an Japanese character, because it | |
| 685 | +belongs to the range~3. This problem is resolved by using |\ltjalchar| | |
| 686 | +instead of the |\char| primitive. It is included in an optional package | |
| 687 | +named \texttt{luatexja-\penalty0fontspec.sty}. Figure~\ref{fig-unitxt} | |
| 688 | +shows several ways o typeset a character , both as a Japanese character | |
| 689 | +and as as an alphabetic characters. | |
| 690 | + | |
| 691 | +\begin{figure} | |
| 692 | +\begin{LTXexample} | |
| 693 | +×, \char`×, % depend on range setting | |
| 694 | +\ltjalchar`×, % alphabetic char | |
| 695 | +\ltjjachar`×, % Japanese char | |
| 696 | +\texttimes % alph. char (by fontspec) | |
| 697 | +\end{LTXexample} | |
| 698 | +\caption{Control sequences producing a Unicode character.} | |
| 699 | +\label{fig-unitxt} | |
| 700 | +\end{figure} | |
| 701 | + | |
| 702 | +The situation looks similar in math formulas, but in fact it differs. | |
| 703 | +Each control sequence that represents an ordinary symbol defined by the | |
| 704 | +\emph{unicode-math} package is just synonym of a character. For example, | |
| 705 | +the meaning of |\otimes| is just the character U+2297 (CIRCLED TIMES), | |
| 706 | +which is included in the range~3. However, it is difficult to define a | |
| 707 | +control sequence like |\ltjalUmathchar| as a counterpart of | |
| 708 | +|\Umathchar|, since an input like `|\sum^\ltjalUmathchar ...|' has to be | |
| 709 | +permitted. | |
| 710 | + | |
| 711 | +However, we couldn't develop a satisfactory solution to this problem in | |
| 712 | +time for this paper, due to a lack of time. We are just testing a | |
| 713 | +solution below: | |
| 714 | +\begin{itemize} | |
| 715 | +\item \LuaTeX-ja has a list of character codes which will be always reated as | |
| 716 | + alphabetic characters in math mode. Considering 8-bit TFMs for | |
| 717 | + math symbols, this list includes natural numbers between |"80| and | |
| 718 | + |"FF| by default. | |
| 719 | +\item Redefine internal commands defined in the \emph{unicode-math} | |
| 720 | + package so that | |
| 721 | +codes of characters which are mentioned in the \emph{unicode-math} | |
| 722 | + package will be included in the list. | |
| 723 | +\end{itemize} | |
| 724 | + | |
| 725 | + | |
| 726 | +We would like to extend treatments described in this subsection to 8-bit | |
| 727 | +font encodings, but we leave it to further development too. | |
| 728 | + | |
| 729 | +\section{Current status of development} | |
| 730 | +\label{sec:current_status} | |
| 731 | +At the moment, \LuaTeX-ja can be used under plain \TeX, and under | |
| 732 | +\LaTeXe. Generally speaking, one only has to read |luatexja.sty|, by | |
| 733 | +|\input| command or |\usepackage| (in~\LaTeXe), if you merely want to | |
| 734 | +typeset Japanese characters. We look more detail by parts. | |
| 735 | + | |
| 736 | +\subsection{`Engine extension'} | |
| 737 | +The lowest part of \LuaTeX-ja corresponds the \pTeX\ extension as | |
| 738 | +\emph{an engine extension of \TeX}. We, the project menbers, think that | |
| 739 | +this part is almost done. There is one more feature of \LuaTeX-ja which | |
| 740 | +we are going to explain: | |
| 741 | + | |
| 742 | +\begin{description} | |
| 743 | +\item[Shifting Baseline] | |
| 744 | +In order to make a match between Japanese fonts and alphabetic fonts, | |
| 745 | + sometimes shifting the baseline of alphabetic characters may | |
| 746 | + be needed. \pTeX\ has a dimension |\ybaselineshift|, which | |
| 747 | + corresponds the amount of shifting down the baseline of alphabetic | |
| 748 | + characters. This is useful for Japanese-based documents, but | |
| 749 | + not for documents mainly in languages with alphabetic | |
| 750 | + characters. | |
| 751 | + | |
| 752 | +Hence, \LuaTeX-ja extends \pTeX's |\ybaselineshift| to Japanese | |
| 753 | + characters. Namely, \LuaTeX-ja offers two parameters, | |
| 754 | + \textsf{yjabaselineshift} and \textsf{yalbaselineshift}, for the | |
| 755 | + amount of shifting the baseline of Japanese characters and | |
| 756 | + that of alphabetic characters, respectively. | |
| 757 | +\begin{figure} | |
| 758 | +\begin{center} | |
| 759 | +\fontsize{40}{40}\selectfont\fboxsep0mm | |
| 760 | +\vrule width 0.9\textwidth height0.4pt depth0.4pt\kern-0.9\textwidth | |
| 761 | +\hbox to 0.9\linewidth{% | |
| 762 | +\hfil | |
| 763 | +\raise-10pt\imagfm{\jstrut 漢}% | |
| 764 | +\raise-10pt\imagfm{\jstrut 字}\hskip.25\zw% | |
| 765 | +\imagfm{p}% | |
| 766 | +\imagfm{h}% | |
| 767 | +\hfil\hfil | |
| 768 | +\imagfm{\jstrut 漢}% | |
| 769 | +\imagfm{\jstrut 字}\hskip.25\zw% | |
| 770 | +\raise-10pt\imagfm{p}% | |
| 771 | +\raise-10pt\imagfm{h}% | |
| 772 | +\hfil | |
| 773 | +} | |
| 774 | +\end{center} | |
| 775 | + | |
| 776 | +\caption{First example of shifting baseline.} | |
| 777 | +\label{fig-bls} | |
| 778 | +\end{figure} | |
| 779 | + | |
| 780 | +\begin{figure} | |
| 781 | +\begin{center} | |
| 782 | +\fontsize{30}{30}\selectfont\fboxsep0mm | |
| 783 | +\vrule width 0.9\textwidth height0.4pt depth0.4pt\kern-0.9\textwidth | |
| 784 | +\hbox to 0.9\linewidth{% | |
| 785 | +\hfil | |
| 786 | +\imagfm{a}% | |
| 787 | +\imagfm{b}\hskip.25\zw% | |
| 788 | +\imagfm{\jstrut 本}% | |
| 789 | +\imagfm{\jstrut 文}\hskip.33333\zw% | |
| 790 | +\raise3.514582pt\imagfm{\fontsize{20}{20}\selectfont\jstrut\inhibitglue （}% | |
| 791 | +\raise3.514582pt\imagfm{\fontsize{20}{20}\selectfont\jstrut 注}% | |
| 792 | +\raise3.514582pt\imagfm{\fontsize{20}{20}\selectfont\jstrut 釈}\hskip.1666667\zw% | |
| 793 | +\raise3.514582pt\imagfm{\fontsize{20}{20}\selectfont c}% | |
| 794 | +\raise3.514582pt\imagfm{\fontsize{20}{20}\selectfont o}% | |
| 795 | +\raise3.514582pt\imagfm{\fontsize{20}{20}\selectfont m}% | |
| 796 | +\raise3.514582pt\imagfm{\fontsize{20}{20}\selectfont m}% | |
| 797 | +\raise3.514582pt\imagfm{\fontsize{20}{20}\selectfont e}% | |
| 798 | +\raise3.514582pt\imagfm{\fontsize{20}{20}\selectfont n}% | |
| 799 | +\raise3.514582pt\imagfm{\fontsize{20}{20}\selectfont t}% | |
| 800 | +\raise3.514582pt\imagfm{\fontsize{20}{20}\selectfont\jstrut ）\inhibitglue}% | |
| 801 | +\hskip.33333\zw% | |
| 802 | +\imagfm{\jstrut 本}% | |
| 803 | +\imagfm{\jstrut 文}% | |
| 804 | +\hfil | |
| 805 | +} | |
| 806 | +\end{center} | |
| 807 | + | |
| 808 | +\caption{Second example of shifting baseline.} | |
| 809 | +\label{fig-small} | |
| 810 | +\end{figure} | |
| 811 | + | |
| 812 | +An example output is shown in Figure~\ref{fig-bls}. The left half is the | |
| 813 | + output when \textsf{yjabaselineshift} is positive, hence the | |
| 814 | + baseline of Japanese characters is shifted down. On the other | |
| 815 | + hand, the right half is the output when | |
| 816 | + \textsf{yalbaselineshift} is positive, hence the baseline of | |
| 817 | + alphabetic characters is shifted down. Figure~\ref{fig-small} | |
| 818 | + shows an intresting use of these parameters. | |
| 819 | + | |
| 820 | +\end{description} | |
| 821 | +Note that \LuaTeX-ja doesn't support vertical typesetting, \emph{tategaki}, for now. | |
| 822 | + | |
| 823 | +\subsection{Patches for plain \TeX\ and \LaTeXe} | |
| 824 | +\pTeX\ has a patch for plain \TeX, namely |ptex.tex|, that for \LaTeXe\ | |
| 825 | +macro (this patch and \LaTeXe\ consist \emph{p\LaTeXe}), and | |
| 826 | +|kinsoku.tex| which includes the default setting of \emph{kinsoku | |
| 827 | +shori}, the Japanese hyphenation. We ported them to \LuaTeX-ja, except | |
| 828 | +the codes related to vertical typesetting, because \LuaTeX-ja doesn't | |
| 829 | +support vertical typesetting yet. We remark one point related to the | |
| 830 | +porting: | |
| 831 | +\begin{description} | |
| 832 | + | |
| 833 | +\item[Behavior of\/ {\tt\char92fontfamily\/}] | |
| 834 | +The control sequence |\fontfamily| in p\LaTeXe\ changes the current alphabetic | |
| 835 | + font family and/or the current Japanese font family, | |
| 836 | + depending the argument. More concretely, | |
| 837 | + |\fontfamily{|$\langle\hbox{\it arg\/}\rangle$|}| changes the | |
| 838 | + current alphabetic font family to $\langle\hbox{\it | |
| 839 | + arg\/}\rangle$, if and only if one of the following | |
| 840 | + conditions are satisfied: | |
| 841 | +\begin{itemize} | |
| 842 | +\item An alphabetic font family named $\langle\hbox{\it arg\/}\rangle$ in | |
| 843 | + \emph{some} alphabetic encoding already defined in the document. | |
| 844 | +\item There exists an alphabetic encoding $\langle\hbox{\it | |
| 845 | + enc\/}\rangle$ already defined in the document such that a font | |
| 846 | + definition file $\langle\hbox{\it enc\/}\rangle\langle\hbox{\it | |
| 847 | + arg\/}\rangle$|.fd| (all lowercase) exists. | |
| 848 | +\end{itemize} | |
| 849 | +The same criterion is used for changing Japanese font family. | |
| 850 | + | |
| 851 | +To work this behavior well, a list of all (alphabetic) encodings defined | |
| 852 | + already in the document is needed. However, since \LuaTeX-ja | |
| 853 | + is loaded as a package, \LuaTeX-ja cannot have this list. | |
| 854 | + Hence \LuaTeX-ja adopted a different approach, namely | |
| 855 | + |\fontfamily{|$\langle\hbox{\it arg\/}\rangle$|}| changes the | |
| 856 | + current alphabetic font family to $\langle\hbox{\it | |
| 857 | + arg\/}\rangle$, if and only if: | |
| 858 | +\begin{itemize} | |
| 859 | +\item An alphabetic font family named $\langle\hbox{\it arg\/}\rangle$ | |
| 860 | + in the current alphabetic encoding $\langle\hbox{\it | |
| 861 | + enc\/}\rangle$ already defined in the document. | |
| 862 | +\item A font definition file $\langle\hbox{\it enc\/}\rangle\langle\hbox{\it | |
| 863 | + arg\/}\rangle$|.fd| (all lowercase) exists. | |
| 864 | +\end{itemize} | |
| 865 | + | |
| 866 | + | |
| 867 | +\end{description} | |
| 868 | + | |
| 869 | + | |
| 870 | + | |
| 871 | +\subsection{Classes for Japanese documents} | |
| 872 | +To produce `high-quality' Japanese documents, we need not only that | |
| 873 | +Japanese characters are correctly placed, but also class files for | |
| 874 | +Japanese documents. Two major families of classes are widely used in Japan: | |
| 875 | +\emph{jclasses} which is distributed with the official p\LaTeXe\ macros, | |
| 876 | +and \emph{jsclasses}. At the present, \LuaTeX-ja | |
| 877 | +simply contains their counterparts: \emph{ltjclasses} and | |
| 878 | +\emph{ltjsclasses}. However, the policy on classes is not determined | |
| 879 | +now, and we hope to have another family of classes which are useful for | |
| 880 | +commercial printing. In the author's opinion, \emph{ltjclasses} is | |
| 881 | +better to stay as an example of porting of class files for \pTeX\ to | |
| 882 | +\LuaTeX-ja. | |
| 883 | + | |
| 884 | +\subsection{Patches for packages} | |
| 885 | +Apart from patches for the \LaTeXe~kernel and classes for Japanese | |
| 886 | +documents, we need to make patches for several packages. At the present, | |
| 887 | +we considered the following packages, and made patches or porting for | |
| 888 | +the former two packages. | |
| 889 | + | |
| 890 | +\begin{description} | |
| 891 | +\item[The \emph{fontspec} package] The \emph{fontspec} package is built | |
| 892 | + on NFSS2, hence control sequences offered by the | |
| 893 | + \emph{fontspec} package, such as |\setmainfont|, are only | |
| 894 | + effective for alphabetic fonts if \LuaTeX-ja is loaded. | |
| 895 | + \texttt{luatexja-\penalty0fontspec.sty} (not automatically | |
| 896 | + loaded) offers these counterparts for Japanese fonts, with | |
| 897 | + additional `j' in the name of control sequences, such as | |
| 898 | + |\setmainjfont|. As described in | |
| 899 | + Subsection~\ref{ssec-unichar}, it also includes a patch for | |
| 900 | + control sequences producing Unicode characters. | |
| 901 | + | |
| 902 | +\item[The \emph{otf} package] | |
| 903 | +This package is widely used in \pTeX\ for typesetting characters which is | |
| 904 | +not in JIS~X~0208, and for using more than one weight in \emph{mincho} | |
| 905 | +and \emph{gothic} font families. Therefore \LuaTeX-ja supports features | |
| 906 | +in the \emph{otf} package, by loading \texttt{luatexja-\penalty0otf.sty} | |
| 907 | + manually. Note that characters by |\UTF{xxxx}| and | |
| 908 | + |\CID{xxxx}| are not appended to the current list as a | |
| 909 | + \emph{glyph\_node}, to avoid from callbacks by the | |
| 910 | + \emph{luaotfload} package. We have another remark; |\CID| | |
| 911 | + does not work with TrueType fonts, since |\CID| use the | |
| 912 | + conversion table between CID and the glyph order of the | |
| 913 | + current Japanese font. | |
| 914 | + | |
| 915 | +\item[The \emph{listings} package] | |
| 916 | +It is known for users of \pTeX\ that there is a patch |jlisting.sty| for | |
| 917 | + the \emph{listings} package, to use Japanese characters in | |
| 918 | + the |lstlisting| environment. Generally speaking, it also can | |
| 919 | + be used in \LuaTeX-ja. However, it seems to be that a | |
| 920 | + Japanese character after a space does not recieve any process | |
| 921 | + of the \emph{listings} package; this is inconvinient when we | |
| 922 | + use the \emph{showexpl} package. | |
| 923 | + | |
| 924 | +There is another way to use characters above 256 with the | |
| 925 | + \emph{listings} package (described in\cite{apl}). However, | |
| 926 | + this method is not suitable for Japanese, since the number of | |
| 927 | + Japanese characters is very large. We hope that the | |
| 928 | + \emph{listings} package will be able to handle all characters above | |
| 929 | + 256 without any patch, in the future. | |
| 930 | + | |
| 931 | + | |
| 932 | +\end{description} | |
| 933 | + | |
| 934 | + | |
| 935 | + | |
| 936 | +\section{Implementation} | |
| 937 | +\label{sec:implementation} | |
| 938 | +\subsection{Handling of Japanese fonts} | |
| 939 | +In \pTeX, there are three slots for maintaining current fonts, namely | |
| 940 | +|\font| for alphabetic fonts, |\jfont| for Japanese fonts (in horizontal | |
| 941 | +direction) and |\tfont| for Japanese fonts (in vertical direction). With | |
| 942 | +these slots, we can manage the current font for alphabetic characters | |
| 943 | +and that for Japanese characters separately in \pTeX. However, \LuaTeX\ | |
| 944 | +has only one slot for maintaining the current font, as \TeX82. This | |
| 945 | +situation leads a problem: how can we maintain the `current Japanese | |
| 946 | +font'? | |
| 947 | + | |
| 948 | +There are three approaches for this problem. One approach is to make a | |
| 949 | +mapping table from alphabetic fonts to corresponding Japanese fonts | |
| 950 | +(here we don't assume that NFSS2 is available). Another approach is | |
| 951 | +that we always use composite fonts with alphabetic fonts and Japanese | |
| 952 | +fonts. The third approach is that the information of the current | |
| 953 | +Japanese font is stored in an attribute. We adopted the third approach, | |
| 954 | +since \LuaTeX-ja is much affected by \pTeX\ as we noted in | |
| 955 | +Subsection~\ref{ssec-pol}. | |
| 956 | + | |
| 957 | +As in Figure~\ref{fig-jfdef}, \LuaTeX-ja uses |\jfont| for defining | |
| 958 | +Japanese font, as \pTeX. However, because the information of the current | |
| 959 | +Japanese font is stored into an attribute, control sequences defined by | |
| 960 | +|\jfont| (e.g.,~|\foo| and |\bar| in Figure~\ref{fig-jfdef}) is | |
| 961 | +not representing a font by the means of \TeX82. In other words, each of | |
| 962 | +these control sequences is just an assignment to an attribute, therefore | |
| 963 | +they cannot be an argument of |\the|, |\fontname|, nor |\textfont|. | |
| 964 | + | |
| 965 | + | |
| 966 | +Callbacks by the \emph{luaotfload} package, e.g.,~replacement of glyphs | |
| 967 | +according to font features, are executed just after `Examination of | |
| 968 | +Stack Level' (see Subsections \ref{ssec-over}~and~\ref{ssec-stack}). Note that calculation of | |
| 969 | +character classes for each Japanese character is done \emph{after} the | |
| 970 | +these callbacks for now. | |
| 971 | + | |
| 972 | +\subsection{Stack management} | |
| 973 | +\label{ssec-stack} | |
| 974 | + | |
| 975 | +As we noted in Subsection~\ref{ssec-csname}, parameters that the values | |
| 976 | +at the end of a horizontal box or that of a paragraph are valid in | |
| 977 | +whole box or paragraph, such as \emph{kanjiskip}, cannot be implemented | |
| 978 | +by internal integers or registers of other types in \TeX. We explain it | |
| 979 | +in this subsection. | |
| 980 | + | |
| 981 | +\begin{figure} | |
| 982 | +\begin{lstlisting} | |
| 983 | +void package(int c) | |
| 984 | +{ | |
| 985 | + ... | |
| 986 | + d = box_max_depth; | |
| 987 | + unsave(); | |
| 988 | + save_ptr -= 4; | |
| 989 | + if (cur_list.mode_field == -hmode) { | |
| 990 | + cur_box = filtered_hpack(cur_list.head_field, | |
| 991 | + cur_list.tail_field, saved_value(1), | |
| 992 | + saved_level(1), grp, saved_level(2)); | |
| 993 | + subtype(cur_box) = HLIST_SUBTYPE_HBOX; | |
| 994 | + } else { | |
| 995 | +\end{lstlisting} | |
| 996 | +\caption{An extract of a CWEB-source \texttt{tex/packaging.w} of \LuaTeX.} | |
| 997 | +\label{fig-ltsrc} | |
| 998 | +\end{figure} | |
| 999 | + | |
| 1000 | +Figure~\ref{fig-ltsrc} is an extract of a CWEB-source | |
| 1001 | +\texttt{tex/packaging.w} of \LuaTeX\ (SVN revision 4358). This function | |
| 1002 | +is called just when an explicit |\hbox{...}| or |\vbox{...}| is ended, and | |
| 1003 | +the function |filtered_hpack()| is where the |hpack_filter| and then the | |
| 1004 | +actual `hpack' process are performed. Notice that the |unsave()| | |
| 1005 | +function is called before |filtered_hpack()|. This is the problem; | |
| 1006 | +because of |unsave()|, we can retrive only the values of registers | |
| 1007 | +\emph{outside} the box, even in the |hpack_filter| callback. | |
| 1008 | + | |
| 1009 | +To cope with this problem, \LuaTeX-ja has its own stack system, based on | |
| 1010 | +Lua codes in \cite{stack-mail}. Furthermore, \emph{whatsit} nodes whose | |
| 1011 | +\emph{user\_id} is 30112 (\emph{stack\_node}, for short) will be | |
| 1012 | +appended to the current horizontal list each time the current stack | |
| 1013 | +level is incremented, and their values are the values of | |
| 1014 | +|\currentgrouplevel| at that time. In the beginning of the |hpack_filter| | |
| 1015 | +callback, the list in question is traversed to determine whether the | |
| 1016 | +stack level at the end of the list and that outside the box coincides. | |
| 1017 | + | |
| 1018 | +Let $x$ be the value of |\currentgrouplevel|, and $y$ be the current | |
| 1019 | +stack level, both inside the |hpack_filter| callback, i.e.,~outside a | |
| 1020 | +horizontal box. Consider a list which represents the content of the box, | |
| 1021 | +then we have: | |
| 1022 | +\begin{itemize} | |
| 1023 | +\item A \emph{stack\_node} whose value is $x+1$ (because all materials in | |
| 1024 | + the box are included in a group |\hbox{...}|, the value is at | |
| 1025 | + least $x+1$) in the list represents an assignment related to the | |
| 1026 | + stack system in just top-level of the list, like | |
| 1027 | +\begin{quote} | |
| 1028 | +\begin{verbatim} | |
| 1029 | +\hbox{...(assignment)...} | |
| 1030 | +\end{verbatim} | |
| 1031 | +\end{quote} | |
| 1032 | +In this case, the current stack level is incremented to $y+1$ after the assignment. | |
| 1033 | +\item A \emph{stack\_node} whose value is more than $x+1$ in the list represents | |
| 1034 | +an assignment inside another group contained in the box. For example, | |
| 1035 | + the following input creates | |
| 1036 | +a \emph{stack\_node} whose value is $x+3=(x+1)+2$: | |
| 1037 | +\begin{quote} | |
| 1038 | +\begin{verbatim} | |
| 1039 | +\hbox{...{...{...(assignment)}...}...} | |
| 1040 | +\end{verbatim} | |
| 1041 | +\end{quote} | |
| 1042 | +\end{itemize} | |
| 1043 | +Thus, we can conclude that the stack level at the end of the list is | |
| 1044 | +$y+1$, if and only if there is a \emph{stack\_node} whose value is | |
| 1045 | +$x+1$. Otherwise, the stack level is just $y$. | |
| 1046 | + | |
| 1047 | +\subsection{Adjustment of the position of Japanese characters} | |
| 1048 | +\label{ssec-width} | |
| 1049 | + | |
| 1050 | +The size of a glyph specified in a metric and that of a real font | |
| 1051 | +usually differ. For example, the letter `\inhibitglue【' is half-width | |
| 1052 | +in |jfm-ujis.lua| or |jis.tfm|, while this letter is full-width like `【' | |
| 1053 | +in most TrueType fonts used in Japanese typesetting, such as | |
| 1054 | +IPA~Mincho. Hence the adjustment of position of such glyphs is | |
| 1055 | +needed. In the context of \pTeX, this process was performed using virtual fonts. | |
| 1056 | + | |
| 1057 | +On the other hand, Lua\TeX-ja does the adjustment by encapsuling a glyph | |
| 1058 | +into a horizontal box. There are two main reasons why we adopted this | |
| 1059 | +method; one is that we feared Lua codes for coexisting with callbacks by | |
| 1060 | +the |luaotfload| package would be large if we use virtual fonts, and the | |
| 1061 | +other is to cope with shifting of the baseline of characters at the | |
| 1062 | +same time. | |
| 1063 | + | |
| 1064 | +\begin{figure} | |
| 1065 | +\begin{center}\unitlength=9pt\small | |
| 1066 | +\begin{picture}(15,12)(-1,-3) | |
| 1067 | + | |
| 1068 | +\color{grayx}% real glyph | |
| 1069 | +\put(-1,-1.5){\vrule width 6\unitlength height 7\unitlength depth 2.5\unitlength} | |
| 1070 | + | |
| 1071 | +\color{black}% real glyph :step1 | |
| 1072 | +\thicklines | |
| 1073 | +\put(-1,-1.5){\line(0,1){7}\line(0,-1){2.5}} | |
| 1074 | +\put(5,-1.5){\line(0,1){7}\line(0,-1){2.5}} | |
| 1075 | +\put(-1,5.5){\line(1,0){6}} | |
| 1076 | +\put(-1,-4){\line(1,0){6}} | |
| 1077 | +\put(-1,0){\makebox(0,0)[r]{\strut$R$\,}} | |
| 1078 | + | |
| 1079 | +\thicklines | |
| 1080 | +\put(0,0){\vector(0,1){9}\line(0,-1){3}\vector(1,0){12}} | |
| 1081 | +\put(12,9){\makebox(0,0)[rt]{\strut$M$\,}} | |
| 1082 | +\put(12,0){\line(0,1){9}\vector(0,-1){3}} | |
| 1083 | +\put(0,9){\line(1,0){12}} | |
| 1084 | +\put(0,-3){\line(1,0){12}} | |
| 1085 | +\put(0.2,4.5){\makebox(0,0)[l]{\texttt{height}}} | |
| 1086 | +\put(12.2,-1.5){\makebox(0,0)[l]{\texttt{depth}}} | |
| 1087 | +\put(6,0.2){\makebox(0,0)[b]{\texttt{width}}} | |
| 1088 | + | |
| 1089 | +\thicklines | |
| 1090 | +\put(3,0){\line(0,1){7}\line(0,-1){2.5}\line(1,0){6}} | |
| 1091 | +\put(9,0){\line(0,1){7}\line(0,-1){2.5}} | |
| 1092 | +\put(3,7){\line(1,0){6}} | |
| 1093 | +\put(3,-2.5){\line(1,0){6}} | |
| 1094 | +\newsavebox{\eqdist} | |
| 1095 | +\savebox{\eqdist}(0,0)[c]{% | |
| 1096 | + \thinlines | |
| 1097 | + \put(-0.08,0.2){\line(0,-1){0.4}}% | |
| 1098 | + \put(0.08,0.2){\line(0,-1){0.4}}} | |
| 1099 | +\put(1.5,0){\usebox{\eqdist}} | |
| 1100 | +\put(10.5,0){\usebox{\eqdist}} | |
| 1101 | + | |
| 1102 | +\thicklines | |
| 1103 | +\put(3,-1.5){\vector(-1,0){4}} | |
| 1104 | +\put(1,-1.7){\makebox(0,0)[t]{\texttt{left}}} | |
| 1105 | +\put(3,0){\vector(0,-1){1.5}} | |
| 1106 | +\put(3.2,-0.75){\makebox(0,0)[l]{\texttt{down}}} | |
| 1107 | +\end{picture} | |
| 1108 | +\end{center} | |
| 1109 | +\caption{The position of the `real' glyph.} | |
| 1110 | +\label{fig-pos} | |
| 1111 | +\end{figure} | |
| 1112 | + | |
| 1113 | +Figure~\ref{fig-pos} shows the adjustment process. A large square $M$ is | |
| 1114 | +the imaginary body specified in the metric, and a vertical | |
| 1115 | +rectangle is the imaginary body of a real glyph. First, the real glyph | |
| 1116 | +is aligned with respect to the width of $M$. In the figure, the real | |
| 1117 | +glyph is aligned `middle'; this setting is useful for the full-width | |
| 1118 | +middle dot `・'. We have other settings, `left' and `right'. | |
| 1119 | +After that, it is shifted according to the value of |left| and |down|, | |
| 1120 | +which are specified in the metric, too. The final position of the real glyph | |
| 1121 | +is shown by the gray rectangle~$R$. If the amount of shifting the baseline is | |
| 1122 | +not zero, $M$ (and hence the real glyph) is shifted by that amount. | |
| 1123 | + | |
| 1124 | +We would like to remark briefly on the vertical position of a real | |
| 1125 | +glyph. A JFM (or a metric used in \LuaTeX-ja) and a real font used for | |
| 1126 | +it may have different height or depth. In that case, it may look better | |
| 1127 | +if the real glyph is shifted vertically to match the height-depth ratio | |
| 1128 | +specified in the metric, while any vertical adjustment except the | |
| 1129 | +adjustment by the |down| value does not performed in the present | |
| 1130 | +implementation of \LuaTeX-ja . This situation is carefully studied by | |
| 1131 | +Otobe~\cite{min10}. Here the policy on this problem is not determined | |
| 1132 | +now, however we would like to offer several solutions in future | |
| 1133 | +development. | |
| 1134 | + | |
| 1135 | +\section{Conclusion} | |
| 1136 | +We have discussed about our \LuaTeX-ja package, which is much affected | |
| 1137 | +by \pTeX. For now, it can be used for experimental use, however there | |
| 1138 | +are much refinements which are needed for regular use. The author hopes | |
| 1139 | +that this paper and \LuaTeX-ja project contribute the typesetting Japanese, | |
| 1140 | +and possibly other Asian languages, under \LuaTeX. | |
| 1141 | + | |
| 1142 | +\section*{Acknowledgements} | |
| 1143 | +The author would like to thank Ken Nakano and Hideaki Togashi for their | |
| 1144 | +development of ASCII \pTeX. The author is very grateful to Haruhiko | |
| 1145 | +Okumura for his leadership in the Japanese \TeX\ community. The author | |
| 1146 | +is also very grateful to members of \LuaTeX-ja project team for their | |
| 1147 | +valuable cooperation in development. | |
| 1148 | + | |
| 1149 | +%%% The style of the bibiliogrphy is `amsplain'. | |
| 1150 | +\providecommand{\bysame}{\leavevmode\hbox to3em{\hrulefill}\thinspace} | |
| 1151 | +\providecommand{\href}[2]{#2} | |
| 1152 | +\begin{thebibliography}{99} | |
| 1153 | + | |
| 1154 | +\bibitem{aj16} | |
| 1155 | +Adobe Systems Incorporated, \emph{Adobe-Japan1-6 Character Collection | |
| 1156 | + for CID-Keyed Fonts}, Technical Note~\#5078, 2004. | |
| 1157 | +\url{http://partners.adobe.com/public/developer/en/font/5078.Adobe-Japan1-6.pdf} | |
| 1158 | + | |
| 1159 | +\bibitem{ptex} | |
| 1160 | +ASCII MEDIA WORKS，アスキー日本語\TeX\ (\pTeX)．\url{http://ascii.asciimw.jp/pb/ptex/} | |
| 1161 | + | |
| 1162 | +\bibitem{apl} | |
| 1163 | +John Baker, \emph{Typesetting UTF8 APL code with the \LaTeX\ lstlisting package}. | |
| 1164 | +\url{http://bakerjd99.wordpress.com/2011/08/15/} | |
| 1165 | + | |
| 1166 | +\bibitem{omega} | |
| 1167 | +Jin-Hwan~Cho and Haruhiko Okumura, \emph{Typesetting CJK Languages with Omega}, | |
| 1168 | +\TeX, XML, and Digital Typography, Lecture Notes in Computer Science, vol.~3130, | |
| 1169 | +Springer, 2004, 139--148. | |
| 1170 | + | |
| 1171 | +\bibitem{joylua} | |
| 1172 | +Yannis Haralambous. \emph{The Joy of \LuaTeX}. \url{http://luatex.bluwiki.com/} | |
| 1173 | + | |
| 1174 | +\bibitem{jisx4051} | |
| 1175 | +Japanese Industrial Standards Committee. \emph{JIS~X~4051: Formatting | |
| 1176 | + rules for Japanese documents}, 1993, 1995, 2004. | |
| 1177 | + | |
| 1178 | +\bibitem{eptex} | |
| 1179 | +北川弘典，$\varepsilon$-\pTeX についてのwiki． | |
| 1180 | +\url{http://sourceforge.jp/projects/eptex/wiki/FrontPage} | |
| 1181 | + | |
| 1182 | +\bibitem{luaums} | |
| 1183 | +北川弘典，\LuaTeX で日本語． | |
| 1184 | +\url{http://oku.edu.mie-u.ac.jp/tex/mod/forum/discuss.php?d=378} | |
| 1185 | + | |
| 1186 | +\bibitem{luatexref} | |
| 1187 | +\LuaTeX\ development team, \emph{The \LuaTeX\ reference}. | |
| 1188 | +\url{http://www.luatex.org/svn/trunk/manual/luatexref-t.pdf} (snapshot of SVN trunk) | |
| 1189 | + | |
| 1190 | +\bibitem{man} | |
| 1191 | +\LuaTeX-ja project team, \emph{The \LuaTeX-ja package}. | |
| 1192 | +Not completed for now. Available at |doc/man-en.pdf| (in English) or | |
| 1193 | + |doc/man-ja.pdf| (in Japanese) | |
| 1194 | +in the Git repository. | |
| 1195 | + | |
| 1196 | +\bibitem{luajp-test} | |
| 1197 | +香田温人，\LuaTeX と日本語． | |
| 1198 | +\url{http://www1.pm.tokushima-u.ac.jp/~kohda/tex/luatex-old.html} | |
| 1199 | + | |
| 1200 | +\bibitem{luajalayout} | |
| 1201 | +前田一貴，luajalayout パッケージ---Lua\LaTeX によ | |
| 1202 | + る日本語組版---． | |
| 1203 | +\url{http://www-is.amp.i.kyoto-u.ac.jp/lab/kmaeda/lualatex/luajalayout/} | |
| 1204 | + | |
| 1205 | +\bibitem{jsclasses} | |
| 1206 | +奥村晴彦，p\LaTeXe 新ドキュメントクラス． | |
| 1207 | +\url{http://oku.edu.mie-u.ac.jp/~okumura/jsclasses/} | |
| 1208 | + | |
| 1209 | +\bibitem{ptexjp} | |
| 1210 | +Haruhiko Okumura, \emph{\pTeX\ and Japanese Typesetting}, | |
| 1211 | + The Asian Journal of \TeX\ \textbf{2}~(2008), 43--51. | |
| 1212 | + | |
| 1213 | +\bibitem{min10} | |
| 1214 | +乙部厳己，min10フォントについて． | |
| 1215 | +\url{http://argent.shinshu-u.ac.jp/~otobe/tex/files/min10.pdf} | |
| 1216 | + | |
| 1217 | +\bibitem{otf} | |
| 1218 | +齋藤修三郎，Open Type Font用VF． | |
| 1219 | +\url{http://psitau.kitunebi.com/otf.html} | |
| 1220 | + | |
| 1221 | +\bibitem{stack-mail} | |
| 1222 | +Jonathan Sauer, \emph{[Dev-luatex] tex.currentgrouplevel}. | |
| 1223 | +\url{http://www.ntg.nl/pipermail/dev-luatex/2008-August/001765.html} | |
| 1224 | + | |
| 1225 | +\bibitem{uptex} | |
| 1226 | +Takuji Tanaka, \emph{u\pTeX, up\LaTeX---unicode version of \pTeX, p\LaTeX}. | |
| 1227 | +\url{http://homepage3.nifty.com/ttk/comp/tex/uptex_en.html} | |
| 1228 | + | |
| 1229 | +\bibitem{ptexenc} | |
| 1230 | +Nobuyuki Tsuchimura, \emph{Development of a Japanese \TeX\ Distribution~`ptetex3'}, | |
| 1231 | +Computer Software\ \textbf{24} (2007), no.~4, 40--50, (in Japanese). | |
| 1232 | + | |
| 1233 | +\bibitem{w3c} | |
| 1234 | +W3C Working Group, \emph{Requirements for Japanese Text Layout}. | |
| 1235 | +\url{http://www.w3.org/TR/jlreq/} | |
| 1236 | +\end{thebibliography} | |
| 1237 | + | |
| 1238 | +\end{document} |
Fork