# hyphenation and utf8

from Tex FAQ - hyphenation I see that trying to help Latex in correctly hyphenating words won't work when the utf8 encoding is used. Is there a work around?

-
Is 'Use LuaTeX or XeTeX' going to be an acceptable answer? – Joseph Wright Apr 27 '11 at 13:17
The web page you are quoting doesn't contain the word "utf8" and actually doesn't say anything about input encodings. Your claim is wrong. – Ulrike Fischer Apr 27 '11 at 13:19
@Ulrike: should I add \usepackage[T1]{fontenc} in my preambule together with \usepackage[utf8]{inputenc}? – pluton Apr 27 '11 at 13:40
@Joseph: I am still a pdflatex user but I may switch sometime in the future. – pluton Apr 27 '11 at 13:41
@pluton: Yes you need T1-encoding. – Ulrike Fischer Apr 27 '11 at 13:51

## 1 Answer

since it may be useful to someone, for specific languages with accents for instance, the \hyphenation command will correctly work only with a T1-encoding \usepackage[T1]{fontenc} but won't without.

-
@pluton: I don't think this is quite correct. The \hyphenation command will work when the font encoding corresponds to what is pre-supposed in the pattern file, and when the hyphenated word uses only accented letters available as single glyphs. Althougth I am not sure what I say is 100% correct, at least the following code shows that your restriction to T1 is not correct: – jfbu Apr 27 '11 at 16:20
\documentclass{article} \usepackage[greek]{babel} \usepackage{gfsbodoni} \usepackage[LGRx,LGR]{fontenc} \hyphenation{\'ab-gd-\>ez} \def\somestuff{abgdez abgdez abgdez abgdez abgdez abgdez abgdez} \def\somestuffwithaccent{\'abgd\>ez \'abgd\>ez \'abgd\>ez \'abgd\>ez \'abgd\>ez \'abgd\>ez \'abgd\>ez} \begin{document} \parbox{2cm}{\somestuff}\hspace{1cm} \parbox{2cm}{\somestuffwithaccent} \end{document} – jfbu Apr 27 '11 at 16:23
I should have said that my example uses lgrxenc.def available from milde.users.sourceforge.net/LGR furthermore, my latex format is not precompiled with the greek patterns, so babel complains in the log, and perhaps if I am extremely unlucky a more complete set-up would hyphenate by default abgdez like I arbitrarely declared for its version with the two accents. – jfbu Apr 27 '11 at 16:27
all existing hyphenation pattern packages on ctan have been translated into utf-8, and xetex and luatex use those translated tables (note that in those processors, the font is unicode-based). inputenc/utf-8 input is translated to some font encoding (e.g., t1, or t2a) that deals with the characters, so hyphenation in this case is completely unaffected by the input coding. – wasteofspace Apr 27 '11 at 16:43
@pluton: You need T1-encoding for your (french) words. The LGR-encoding mentioned by jfbu is for greek and not relevant for you. – Ulrike Fischer Apr 28 '11 at 8:15