Modifying hyphenation algorithm

Build 1501 on 14/Nov/2017  This topic last edited on: 25/Aug/2014, at 12:51

A GN4 Systems Administrator may want to modify the hyphenation algorithm for specific language.

The Liang algorithm uses a set of patterns, that are strings including control values. for each pair of letters a control value can be specified, a control value can also be specified for the beginning and the end of the pattern.

hmtoggle_plus1How the algorithm works?
a.the char ('.') is added at the start and the end of the word to hyphenate.
b.it extracts all the patterns contained in the word.
c.for each pair of letters a set of control values exists, that are i.e. the control values of all the patterns containing the couple. the maximum is computed, and on each point where the maximum is an odd number, the word can be hyphenated.
hmtoggle_plus1EXAMPLES

Word:

casa -> .casa.

patterns contained in the word:

 

.  c  a  s  a  .

.c  as  sa  a.

.ca  cas  asa  sa.

.cas  casa  asa.

.casa  casa.

.casa.

patterns contained in the word that are specified in the italian CPATTN file:

as -> 2a1s

maximum control value for each position:

. c a s a .

0 2 1 0 0

the word can be hyphenated: ca-sa because the only odd value (1) is found between the 'a' and the 's'.

 

Word:

arto -> .arto.

patterns contained in the word that are specified in the italian CPATTN file:

ar -> 2a1r

rt -> 2r3t4

maximum control value for each position:

.2a2r3t4o0.

the word can be hyphenated:  ar-to.

 

Word:

noveaux -> .noveaux.

patterns contained in the word that are specified in the french CPATTN file:

ov -> 2o1v

ea -> e1a

eaux. -> e4a4u4x.

maximum control value for each position:

.0n2o1v0e4a4u4x.

the word can be hyphenated: no-veaux.

For more information about the Liang algorithm refer to 'The Texbook' - Donald E. Knuth - Addison Wesley - Appendix H.

See also

Loading hyphenation in database