=== Abstract ===

This library provides an implementation of Frank Liangs hyphenation algorithm,
better known as the TeX hyphenation algorithm, for C++ and C.

=== Dependencies ===

STL, POSIX and the GNOME glib.

=== How to use ===

An example program is supplied as example.cpp in this directory.

=== Hyphenation files ===

In order to hyphenate text, one must supply knowledge about the language
used. Unfortunately, both the well-established tex hyphenation dictionaries
and the libhnj dictionaries used by Star Office lack one crucial bit of 
information, namely how many characters have to be left unhyphenated at
the start and end of each word. This breaks, for example, the german language
patterns, which might be the language needing hyphenation most.

Thus, we have to introduce yet another format (sorry for that), but I've tried
to keep it as close to the established one as possible. I've precompiled the
hyphenation patterns for the major languages I know about (english, french,
german, spanish, italian).

Luckily, the process of converting a TeX hyphenation dictionary for your own
language is quite simple.
In order to compile another language pattern, the TeX hyphenation file
(usually residing in /usr/share/texmf-tetex/tex/generic/hyphen/) is a good
start. That TeX file's main body is a huge number of words given as an
argument to the \pattern command; you'll see word sniplets there intermingled
with digits. In order to use them with libhyphenate, you'll have to

a) snip all but the arguments of the \pattern command,
b) expand all commands remaining in that argument string,
c) convert all characters to UTF-8 and
d) prefix the resulting patterns with a line giving two numbers separated by
   a space. The first one of these numbers is the minimum number of characters 
   at the start of each word before hyphenation occurs; the second is the 
   minimum number of characters at the end that won't be hyphenated any more.

=== Syntax of the hyphenation files ===

Each hyphenation file is a list of whitespace-separated words. The first two
words give the start-of-word and end-of-word hyphenation skip in decimal 
numbers; all following words are hyphenation patterns, where a . marks the 
start of the word and each number gives an hyphenation hint for the position
it is at. 

Example: The file
2 2
.and2uin.
d1u

will result in hyphenation between each d and u, but not in the word anduin.
For more info, consult Frank Liang: Word Hy-phen-a-tion by Com-pu-ter.

=== Further Documentation ===

The code is documented extensively.

=== License ===

This package is licensed under the LGPL. Refer to the headings in each source
file for more information.