TeX is a program system for typesetting, and it broke many
new grounds in computer typesetting, in particular the
paragraph filling and the hyphenation algorithms developed
for TeX are still the best for their respective purposes.
Unfortunately, this is not so. While TeX's line breaking algorithm
is good, TeX's hyphenation algorithm (Liang's algorithm) is
a rather simple approach imposing serious limitations on proper
hyphenation of German, for example.
The most important advantage of Liang's algorithm is that it is
fast and simple. That's why many implementations of XSL FO use it.
XSLFO leaves many details for the implementation, sometimes
this is made explicit, often this can be guessed from the
absence of any advice. Hyphenation is explicitely left to
the implementation, line breaking implicitely (they could
have mentioned TR14 though, given the extensive reference
to TR9).
UA14 for line-breaking and Liang's for computing hyphenation
points has little to do with the actual line-breaking algorithm,
that is with optimal distribution of syllables between lines in
a paragraph.
Hyphenation dictionaries and UA14 specify where it is allowed to
break a word or a line.
A line-breaking algorithm is a way to determine the actual breakdown.
While it should be based on something like hyphenation tables and UA14,
it is much more than that.
When TeX's line breaking algorithm is mentioned, it is the sophistication
of optimization that is praised, not conformance to a standard. Actually,
TeX knows nothing about Unicode and UA14.
David Tolpin
XSL-List info and archive: http://www.mulberrytech.com/xsl/xsl-list