From cac11a2105e5c8aec24d45acb91d5f415bbaec52 Mon Sep 17 00:00:00 2001 From: Karl Berry Date: Thu, 24 Dec 2009 00:55:36 +0000 Subject: patgen2-tutorial now on its own and gpld per yannis (23dec09) git-svn-id: svn://tug.org/texlive/trunk@16490 c570f23f-e606-0410-a88d-b1316a301751 --- .../texmf-dist/doc/support/patgen/patgen2.tutorial | 331 -------------------- .../doc/support/patgen2-tutorial/patgen2.pdf | Bin 0 -> 69266 bytes .../doc/support/patgen2-tutorial/patgen2.tex | 334 +++++++++++++++++++++ 3 files changed, 334 insertions(+), 331 deletions(-) delete mode 100644 Master/texmf-dist/doc/support/patgen/patgen2.tutorial create mode 100644 Master/texmf-dist/doc/support/patgen2-tutorial/patgen2.pdf create mode 100644 Master/texmf-dist/doc/support/patgen2-tutorial/patgen2.tex (limited to 'Master/texmf-dist/doc/support') diff --git a/Master/texmf-dist/doc/support/patgen/patgen2.tutorial b/Master/texmf-dist/doc/support/patgen/patgen2.tutorial deleted file mode 100644 index 5194083aa2c..00000000000 --- a/Master/texmf-dist/doc/support/patgen/patgen2.tutorial +++ /dev/null @@ -1,331 +0,0 @@ -%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% -% % -% A SMALL TUTORIAL ON THE MULTILINGUAL FEATURES OF PATGEN2 % -% % -%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% -\documentstyle{article} -\title{A small tutorial on the multilingual features of PatGen2} -\author{Yannis Haralambous} -\date{} -\begin{document} -\maketitle - -\def\XeT{X\kern-.1667em\lower.5ex\hbox{E}\kern-.125emT} - -I will very briefly discuss and illustrate by an example the features -of PatGen2\footnote{the extension to PatGen by Peter Breitenlohner, -author of many beautiful and extremely useful \TeX ware, such as -\TeX--\XeT, {\tt DVIcopy} etc.}, related to the {\em translation} -file. - -\section{Syntax of the translation file} -\subsection{What is the problem} -The problem is that in our (non-English, non-Latin, -non-Indonesian\footnote{As far as I know, English, Latin and -Indonesian are the only languages without diacritics or special -characters...}) languages, we have more characters than just the~26 -letters of the Latin alphabet. Not to mention that alphabetical orders -can be quite different (remember: in Spanish, cucaracha comes before -chacal), and that there can be many ways to express these additional -characters: for example the French \oe{} (like in << Histoire de l'\oe -il >> by Georges Bataille) can be given to \TeX\ as an 8-bit -character, or as \verb*=\oe =, or as \verb=\oe{}=, or as \verb=^^f7= -(since we are {\em all} now working with DC fonts) and so forth. - -Now let's make (or enhance the) hyphenation patterns for our language(s). -Fred Liang has not provided PatGen with the extensibility feature: before -PatGen2, if your language had less than~26 letters you could hack patterns by -substituting letters, if it had more\ldots then the odds were against you. - -So the problem is to express additional characters in the patterns, in some -understandable way (and if possible to get the results in the appropriate -alphabetical order). - -\subsection{The solution} -PatGen2 allows you to include more characters than the~26 letters of -the alphabet. You can specify arbitrary many {\em input forms} for -each of them (these forms will be identified internally), and you can -specify the alphabetical order of your language. The latter feature is -only interesting when you want to study the patterns file to correct -bugs, it doesn't affect the result. - -These informations are transmitted to PatGen2 by means of a {\em -translation} file (usually with the extension \verb=.tra=). The syntax -of this file is not very user-friendly, but if you deal with PatGen -you are supposed to be a hacker anyway, and hackers just {\bf love} -weird-but-efficient-syntaxes [the proof: we all like -\TeX!\footnote{\ldots just joking}]. - -On the first line you specify the \verb=\lefthyphenmin= and -\verb=\righthyphenmin= values. The seven first positions of this line -are used as follows: two positions for \verb=\lefthyphenmin=, two -positions for \verb=\righthyphenmin=, and one position for substitute -symbols of each of \verb=.=, \verb=-=, \verb=*= in the output file. -This can be useful if you are going to use the dot the asterisk or the -hyphen to describe your additional characters (for example one may use -the dot to specify the dot accent on a letter). - -The next lines concern letters of the alphabet of our language. The -first position of each line is the delimiter of our fields. You know -from languages like C or {\tt awk} that it is very useful to define -our own delimiter (instead of the blank space) if for any reason we -want to {\em include} a blank space into our fields. This will be the -case in the example: one usually leaves a blank space after a \TeX\ -macro without argument. - -Otherwise just leave the first position blank. Next comes the -standard representation of our character (preferably lowercase). This -is how the character will appear in the patterns and in the hyphenated -output file. - -Follows again a delimiter, and all the {\em equivalent} -representations of your character: its uppercase form, and as many -other input forms we wish, always followed by a delimiter. - -The order of lines specifies the alphabetical order of your characters. - -When PatGen sees two consecutive delimiters, -it stops reading; so we can include -comments after that. - -\section{An example} -To illustrate this I have taken some -Greek words (people who know $X\acute\alpha% -\rho\rho\upsilon$ $K\lambda\acute\upsilon\nu\nu$ -and his album $\Pi\alpha\tau% -\acute\alpha\tau\epsilon\varsigma$ will recognize some of these -words\ldots) and included an $\alpha$ with accent, and a variant -representation of $\pi$, in form of a macro \verb*=\varpi =. Also I -followed the Greek alphabetical order in the translation file. - -Here is my list of words: - -\begin{verbatim} -A-NU-PO-TA-QTOS -A-KA-TA-M'A-QH-TOS -A-EI-MNH-STOS -MA-NA -TOUR-KO-GU-FTIS-SA -NU-QO-KO-PTHS -LI-ME-NO-FU-LA-KAS -MPRE-LOK -A-GA-\varpi A-EI -PEI-RAI-'AS -\end{verbatim} - -(file {\tt greek.dic}); and here is my translation file ({\tt greek.tra}): - -\begin{verbatim} - 1 1 - a A - 'a 'A - b B - g G - d D - e E - z Z - h H - j J - i I - k K - l L - m M - n N - o O -#p#P#\varpi ## - r R - s S - t T - u U - f F - q Q - y Y - w W -\end{verbatim} - -(yes, yes, that's not a joke; Greek really uses values~1 -and~1 for minimal right -and left hyphenations!). As you see, I have chosen \verb=#= as delimiter -in the case of $\pi$, because \verb=\varpi= is supposed to be followed -by a blank space. And then I wrote two delimiters (\verb=##=) to show PatGen2 -that I finished talking about $\pi$. - -Here is what I got on my console: - -\begin{verbatim} -PatGen -t greek.tra -o greek.out greek.dic -This is PatGen, C Version 2.1 / Macintosh Version 2.0 -Copyright (c) 1991-93 by Wilfried Ricken. All rights reserved. -left_hyphen_min = 1, right_hyphen_min = 1, 24 letters -0 patterns read in -pattern trie has 266 nodes, trie_max = 290, 0 outputs -hyph_start: 1 -hyph_finish: 2 -pat_start: 2 -pat_finish: 4 -good weight: 1 -bad weight: 1 -threshold: 1 -processing dictionary with pat_len = 2, pat_dot = 1 - -0 good, 0 bad, 31 missed -0.00 %, 0.00 %, 100.00 % -69 patterns, 325 nodes in count trie, triec_max = 440 -25 good and 42 bad patterns added (more to come) -finding 29 good and 0 bad hyphens, efficiency = 1.16 -pattern trie has 333 nodes, trie_max = 349, 2 outputs -processing dictionary with pat_len = 2, pat_dot = 0 - -29 good, 0 bad, 2 missed -93.55 %, 0.00 %, 6.45 % -45 patterns, 301 nodes in count trie, triec_max = 378 -2 good and 43 bad patterns added -finding 2 good and 0 bad hyphens, efficiency = 1.00 -pattern trie has 339 nodes, trie_max = 386, 6 outputs -processing dictionary with pat_len = 2, pat_dot = 2 - -31 good, 0 bad, 0 missed -100.00 %, 0.00 %, 0.00 % -47 patterns, 303 nodes in count trie, triec_max = 369 -0 good and 47 bad patterns added -finding 0 good and 0 bad hyphens -pattern trie has 344 nodes, trie_max = 386, 13 outputs -51 nodes and 11 outputs deleted -total of 27 patterns at hyph_level 1 - -pat_start: 2 -pat_finish: 4 -good weight: 1 -bad weight: 1 -threshold: 1 -processing dictionary with pat_len = 2, pat_dot = 1 - -31 good, 0 bad, 0 missed -100.00 %, 0.00 %, 0.00 % -27 patterns, 283 nodes in count trie, triec_max = 315 -0 good and 27 bad patterns added -finding 0 good and 0 bad hyphens -pattern trie has 295 nodes, trie_max = 386, 4 outputs -processing dictionary with pat_len = 2, pat_dot = 0 - -31 good, 0 bad, 0 missed -100.00 %, 0.00 %, 0.00 % -27 patterns, 283 nodes in count trie, triec_max = 303 -0 good and 27 bad patterns added -finding 0 good and 0 bad hyphens -pattern trie has 320 nodes, trie_max = 386, 6 outputs -processing dictionary with pat_len = 2, pat_dot = 2 - -31 good, 0 bad, 0 missed -100.00 %, 0.00 %, 0.00 % -24 patterns, 280 nodes in count trie, triec_max = 286 -0 good and 24 bad patterns added -finding 0 good and 0 bad hyphens -pattern trie has 328 nodes, trie_max = 386, 9 outputs -35 nodes and 7 outputs deleted -total of 0 patterns at hyph_level 2 -hyphenate word list? y -writing PatTmp.2 - -31 good, 0 bad, 0 missed -100.00 %, 0.00 %, 0.00 % -Time elapsed: 0:37:70 minutes. -\end{verbatim} - -and here are the results: first of all the patterns -(file {\tt greek.out}) - -\begin{verbatim} -a1g -a1e -a1k -a1m -a1n -a1p -a1t -a1q -'a1q -e1l -e1n -h1t -i1'a -i1m -i1r -1ko -o1g -o1p -o1t -o1f -r1k -s1s -1st -u1l -u1p -u1f -u1q -\end{verbatim} - -As you see, {\tt o1t} comes before {\tt o1f}: (smile) simply -because we are talking about $o1\tau$ and $o1\phi$. So the -alphabetical order is well respected. Also you see that PatGen2 -has read the word \verb*=A-GA-\varpi A-EI= exactly as if it -were \verb*=A-GA-PA-EI= and has made out of it the pattern -{\tt a1p} (if you look in the input words there is no other -reason for this pattern to exist). - -And here is our result (file {\tt PatTmp.2}): - -\begin{verbatim} -a*nu*po*ta*qtos -a*ka*ta*m'a*qh*tos -a*ei*mnh*stos -ma*na -tour*ko*gu*ftis*sa -nu*qo*ko*pths -li*me*no*fu*la*kas -mpre*lok -a*ga*pa*ei -pei*rai*'as -\end{verbatim} - -(Of course, it is correct; you think I would have shown it if it -weren't correct?) As you see there is no \verb*=\varpi = anymore: -PatGen2 has really replaced it by \verb=p=, and so -\verb*=A-GA-\varpi A-EI= has become \verb=a*ga*pa*ei=. - -\section{Where do I find more information?} -I voluntarily didn't discussed the various parameters used for -pattern generation. For these there is very good litterature: -\begin{itemize} -\item the \TeX book, by the Grand Wizard of \TeX\ arcana, -appendix H; -\item \LaTeX\ Erweiterungsm\"oglichkeiten, by Helmut Kopka, -Addison-Wesley, Pages 482--489; -\item Swedish Hyphenation for \TeX, by Jan Michael Rynning, -[sorry, I don't know where this paper is published]; -\item Word Hy-phen-a-tion by Com-put-er, Stanford University -Report {\tt STAN-CS-83-977}; -\item forthcoming paper by Dominik Wujastyk, on British -hyphenation; -\item Hyphenation Patterns for Ancient Greek and Latin, -TUGboat~13 (4), pages 457--469. -\item and many others\ldots -\end{itemize} - -Where to get PatGen2? probably everywhere, but certainly in -Stuttgart ({\tt IP 129.69.1.12}), -\verb=soft/tex/systems/pc/utilities/patgen.zip= (take a look also at -\verb=soft/tex/systems/knuth/texware/patgen.version2.1/patgen.README=). - -\section{Go forth, etc etc} -OK, once again {\sc Go Forth} and make masterpieces of hyphenation -patterns\ldots -{\bf but} please get in touch with the TWGMLC (Technical Working -Group on Multiple Language Coordination) first, since people there -are working on many languages: maybe they have already done what -you need and are still testing it; or maybe they haven't and in -that case you could help us a lot. - - -\end{document} - diff --git a/Master/texmf-dist/doc/support/patgen2-tutorial/patgen2.pdf b/Master/texmf-dist/doc/support/patgen2-tutorial/patgen2.pdf new file mode 100644 index 00000000000..a75dd3b03ac Binary files /dev/null and b/Master/texmf-dist/doc/support/patgen2-tutorial/patgen2.pdf differ diff --git a/Master/texmf-dist/doc/support/patgen2-tutorial/patgen2.tex b/Master/texmf-dist/doc/support/patgen2-tutorial/patgen2.tex new file mode 100644 index 00000000000..2c6f5dd30c8 --- /dev/null +++ b/Master/texmf-dist/doc/support/patgen2-tutorial/patgen2.tex @@ -0,0 +1,334 @@ +%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% +% % +% A SMALL TUTORIAL ON THE MULTILINGUAL FEATURES OF PATGEN2 % +% % +%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% +\documentstyle{article} +\title{A small tutorial on the multilingual features of PatGen2} +\author{Yannis Haralambous} +\date{} +\begin{document} +\maketitle + +\def\XeT{X\kern-.1667em\lower.5ex\hbox{E}\kern-.125emT} + +(This document is released under the GNU General Public License, +version~2 or any later version.) + +I will very briefly discuss and illustrate by an example the features +of PatGen2\footnote{the extension to PatGen by Peter Breitenlohner, +author of many beautiful and extremely useful \TeX ware, such as +\TeX--\XeT, {\tt DVIcopy} etc.}, related to the {\em translation} +file. + +\section{Syntax of the translation file} +\subsection{What is the problem} +The problem is that in our (non-English, non-Latin, +non-Indonesian\footnote{As far as I know, English, Latin and +Indonesian are the only languages without diacritics or special +characters...}) languages, we have more characters than just the~26 +letters of the Latin alphabet. Not to mention that alphabetical orders +can be quite different (remember: in Spanish, cucaracha comes before +chacal), and that there can be many ways to express these additional +characters: for example the French \oe{} (like in << Histoire de l'\oe +il >> by Georges Bataille) can be given to \TeX\ as an 8-bit +character, or as \verb*=\oe =, or as \verb=\oe{}=, or as \verb=^^f7= +(since we are {\em all} now working with DC fonts) and so forth. + +Now let's make (or enhance the) hyphenation patterns for our language(s). +Fred Liang has not provided PatGen with the extensibility feature: before +PatGen2, if your language had less than~26 letters you could hack patterns by +substituting letters, if it had more\ldots then the odds were against you. + +So the problem is to express additional characters in the patterns, in some +understandable way (and if possible to get the results in the appropriate +alphabetical order). + +\subsection{The solution} +PatGen2 allows you to include more characters than the~26 letters of +the alphabet. You can specify arbitrary many {\em input forms} for +each of them (these forms will be identified internally), and you can +specify the alphabetical order of your language. The latter feature is +only interesting when you want to study the patterns file to correct +bugs, it doesn't affect the result. + +These informations are transmitted to PatGen2 by means of a {\em +translation} file (usually with the extension \verb=.tra=). The syntax +of this file is not very user-friendly, but if you deal with PatGen +you are supposed to be a hacker anyway, and hackers just {\bf love} +weird-but-efficient-syntaxes [the proof: we all like +\TeX!\footnote{\ldots just joking}]. + +On the first line you specify the \verb=\lefthyphenmin= and +\verb=\righthyphenmin= values. The seven first positions of this line +are used as follows: two positions for \verb=\lefthyphenmin=, two +positions for \verb=\righthyphenmin=, and one position for substitute +symbols of each of \verb=.=, \verb=-=, \verb=*= in the output file. +This can be useful if you are going to use the dot the asterisk or the +hyphen to describe your additional characters (for example one may use +the dot to specify the dot accent on a letter). + +The next lines concern letters of the alphabet of our language. The +first position of each line is the delimiter of our fields. You know +from languages like C or {\tt awk} that it is very useful to define +our own delimiter (instead of the blank space) if for any reason we +want to {\em include} a blank space into our fields. This will be the +case in the example: one usually leaves a blank space after a \TeX\ +macro without argument. + +Otherwise just leave the first position blank. Next comes the +standard representation of our character (preferably lowercase). This +is how the character will appear in the patterns and in the hyphenated +output file. + +Follows again a delimiter, and all the {\em equivalent} +representations of your character: its uppercase form, and as many +other input forms we wish, always followed by a delimiter. + +The order of lines specifies the alphabetical order of your characters. + +When PatGen sees two consecutive delimiters, +it stops reading; so we can include +comments after that. + +\section{An example} +To illustrate this I have taken some +Greek words (people who know $X\acute\alpha% +\rho\rho\upsilon$ $K\lambda\acute\upsilon\nu\nu$ +and his album $\Pi\alpha\tau% +\acute\alpha\tau\epsilon\varsigma$ will recognize some of these +words\ldots) and included an $\alpha$ with accent, and a variant +representation of $\pi$, in form of a macro \verb*=\varpi =. Also I +followed the Greek alphabetical order in the translation file. + +Here is my list of words: + +\begin{verbatim} +A-NU-PO-TA-QTOS +A-KA-TA-M'A-QH-TOS +A-EI-MNH-STOS +MA-NA +TOUR-KO-GU-FTIS-SA +NU-QO-KO-PTHS +LI-ME-NO-FU-LA-KAS +MPRE-LOK +A-GA-\varpi A-EI +PEI-RAI-'AS +\end{verbatim} + +(file {\tt greek.dic}); and here is my translation file ({\tt greek.tra}): + +\begin{verbatim} + 1 1 + a A + 'a 'A + b B + g G + d D + e E + z Z + h H + j J + i I + k K + l L + m M + n N + o O +#p#P#\varpi ## + r R + s S + t T + u U + f F + q Q + y Y + w W +\end{verbatim} + +(yes, yes, that's not a joke; Greek really uses values~1 +and~1 for minimal right +and left hyphenations!). As you see, I have chosen \verb=#= as delimiter +in the case of $\pi$, because \verb=\varpi= is supposed to be followed +by a blank space. And then I wrote two delimiters (\verb=##=) to show PatGen2 +that I finished talking about $\pi$. + +Here is what I got on my console: + +\begin{verbatim} +PatGen -t greek.tra -o greek.out greek.dic +This is PatGen, C Version 2.1 / Macintosh Version 2.0 +Copyright (c) 1991-93 by Wilfried Ricken. All rights reserved. +left_hyphen_min = 1, right_hyphen_min = 1, 24 letters +0 patterns read in +pattern trie has 266 nodes, trie_max = 290, 0 outputs +hyph_start: 1 +hyph_finish: 2 +pat_start: 2 +pat_finish: 4 +good weight: 1 +bad weight: 1 +threshold: 1 +processing dictionary with pat_len = 2, pat_dot = 1 + +0 good, 0 bad, 31 missed +0.00 %, 0.00 %, 100.00 % +69 patterns, 325 nodes in count trie, triec_max = 440 +25 good and 42 bad patterns added (more to come) +finding 29 good and 0 bad hyphens, efficiency = 1.16 +pattern trie has 333 nodes, trie_max = 349, 2 outputs +processing dictionary with pat_len = 2, pat_dot = 0 + +29 good, 0 bad, 2 missed +93.55 %, 0.00 %, 6.45 % +45 patterns, 301 nodes in count trie, triec_max = 378 +2 good and 43 bad patterns added +finding 2 good and 0 bad hyphens, efficiency = 1.00 +pattern trie has 339 nodes, trie_max = 386, 6 outputs +processing dictionary with pat_len = 2, pat_dot = 2 + +31 good, 0 bad, 0 missed +100.00 %, 0.00 %, 0.00 % +47 patterns, 303 nodes in count trie, triec_max = 369 +0 good and 47 bad patterns added +finding 0 good and 0 bad hyphens +pattern trie has 344 nodes, trie_max = 386, 13 outputs +51 nodes and 11 outputs deleted +total of 27 patterns at hyph_level 1 + +pat_start: 2 +pat_finish: 4 +good weight: 1 +bad weight: 1 +threshold: 1 +processing dictionary with pat_len = 2, pat_dot = 1 + +31 good, 0 bad, 0 missed +100.00 %, 0.00 %, 0.00 % +27 patterns, 283 nodes in count trie, triec_max = 315 +0 good and 27 bad patterns added +finding 0 good and 0 bad hyphens +pattern trie has 295 nodes, trie_max = 386, 4 outputs +processing dictionary with pat_len = 2, pat_dot = 0 + +31 good, 0 bad, 0 missed +100.00 %, 0.00 %, 0.00 % +27 patterns, 283 nodes in count trie, triec_max = 303 +0 good and 27 bad patterns added +finding 0 good and 0 bad hyphens +pattern trie has 320 nodes, trie_max = 386, 6 outputs +processing dictionary with pat_len = 2, pat_dot = 2 + +31 good, 0 bad, 0 missed +100.00 %, 0.00 %, 0.00 % +24 patterns, 280 nodes in count trie, triec_max = 286 +0 good and 24 bad patterns added +finding 0 good and 0 bad hyphens +pattern trie has 328 nodes, trie_max = 386, 9 outputs +35 nodes and 7 outputs deleted +total of 0 patterns at hyph_level 2 +hyphenate word list? y +writing PatTmp.2 + +31 good, 0 bad, 0 missed +100.00 %, 0.00 %, 0.00 % +Time elapsed: 0:37:70 minutes. +\end{verbatim} + +and here are the results: first of all the patterns +(file {\tt greek.out}) + +\begin{verbatim} +a1g +a1e +a1k +a1m +a1n +a1p +a1t +a1q +'a1q +e1l +e1n +h1t +i1'a +i1m +i1r +1ko +o1g +o1p +o1t +o1f +r1k +s1s +1st +u1l +u1p +u1f +u1q +\end{verbatim} + +As you see, {\tt o1t} comes before {\tt o1f}: (smile) simply +because we are talking about $o1\tau$ and $o1\phi$. So the +alphabetical order is well respected. Also you see that PatGen2 +has read the word \verb*=A-GA-\varpi A-EI= exactly as if it +were \verb*=A-GA-PA-EI= and has made out of it the pattern +{\tt a1p} (if you look in the input words there is no other +reason for this pattern to exist). + +And here is our result (file {\tt PatTmp.2}): + +\begin{verbatim} +a*nu*po*ta*qtos +a*ka*ta*m'a*qh*tos +a*ei*mnh*stos +ma*na +tour*ko*gu*ftis*sa +nu*qo*ko*pths +li*me*no*fu*la*kas +mpre*lok +a*ga*pa*ei +pei*rai*'as +\end{verbatim} + +(Of course, it is correct; you think I would have shown it if it +weren't correct?) As you see there is no \verb*=\varpi = anymore: +PatGen2 has really replaced it by \verb=p=, and so +\verb*=A-GA-\varpi A-EI= has become \verb=a*ga*pa*ei=. + +\section{Where do I find more information?} +I voluntarily didn't discussed the various parameters used for +pattern generation. For these there is very good litterature: +\begin{itemize} +\item the \TeX book, by the Grand Wizard of \TeX\ arcana, +appendix H; +\item \LaTeX\ Erweiterungsm\"oglichkeiten, by Helmut Kopka, +Addison-Wesley, Pages 482--489; +\item Swedish Hyphenation for \TeX, by Jan Michael Rynning, +[sorry, I don't know where this paper is published]; +\item Word Hy-phen-a-tion by Com-put-er, Stanford University +Report {\tt STAN-CS-83-977}; +\item forthcoming paper by Dominik Wujastyk, on British +hyphenation; +\item Hyphenation Patterns for Ancient Greek and Latin, +TUGboat~13 (4), pages 457--469. +\item and many others\ldots +\end{itemize} + +Where to get PatGen2? probably everywhere, but certainly in +Stuttgart ({\tt IP 129.69.1.12}), +\verb=soft/tex/systems/pc/utilities/patgen.zip= (take a look also at +\verb=soft/tex/systems/knuth/texware/patgen.version2.1/patgen.README=). + +\section{Go forth, etc etc} +OK, once again {\sc Go Forth} and make masterpieces of hyphenation +patterns\ldots +{\bf but} please get in touch with the TWGMLC (Technical Working +Group on Multiple Language Coordination) first, since people there +are working on many languages: maybe they have already done what +you need and are still testing it; or maybe they haven't and in +that case you could help us a lot. + + +\end{document} + -- cgit v1.2.3