I'm also seeng wrong hyphenations such as "her-oica" and "peq-ueña" in Spanish (one should never hyphenate between "r" and a vowel or between "q" and "u").
If I manually change <body> to <body xml:lang="de">, hyphenation is correct on all accounts. IIRC there's no way to do this in CSS, so I either need to get my html generator (docutils) to do this for me or I postprocess to add it in.
Thanks for taking the effort of looking into it, even if it was a false alarm.
Nonetheless, this seems to be a regression from prince 6.0 where this wasn't an issue (for several small-run print publications with very picky audience).
I'm also seeng wrong hyphenations such as "her-oica" and "peq-ueña" in Spanish (one should never hyphenate between "r" and a vowel or between "q" and "u").
I've found many more errors, all of them seem to be due to a wrong hyphenation point before two consecutive vowels:
ext-iende, seg-uirle, rec-uerdo... all of these are wrong. (Actually, breaking a consonant-vowel combination is very rarely correct).
I am trying to use the latest hyphenation from OpenOffice, but without success. Swiss-German hyphenation is a variant of German hyphenation. For example we have no "strong letter s" in Switzerland leading to different hyphenation rules.
I use in my .css-file prince-hyphenate-patterns: url(../Hyphenate/hyph_de_CH.dic) and I have placed the downloaded hyph_de_CH.dic in a local directory called Hyphenate. I downloaded from http://extensions.openoffice.org/en/project/german-de-ch-frami-dictionaries and I copied the following file from inside the oxt-zip archive: dict-de_ch-frami_2013-12-06\hyph_de_CH\hyph_de_CH.dic
I also have lang="de" in the html-tag of the document. I also tried with "de-CH".
But I get awful hyphenations. So I must make a mistake.
Do you have any idea what mistake? (Am I right in thinking that the path to the .dic-File must be relative to the .css-File?).
I think I found the de-CH hyphenation data in the right format, although the 1901 version instead of the 1996 version. I have to find out if the 1901-1996-difference is mainly in the spelling and not so much in the hyphenation...
It seems there are still some hyphenation bugs in Prince9. We had an actual problem with the word "Dateipfad" which displays as "Dateip-fad" (wrong). Using the method described above, I get "Da·tei·p·fad". The correct hyphenation in this case should be "Datei-pfad"
Is there a way to correct single cases without using a complete external library like LibreOffice?
We are working on improved hyphenation for German, which will be available in future releases.
For now, you can specify the hyphenation points in the word manually using soft-hyphen characters, or you can add some new patterns to the hyphenation dictionary (hyph/hyph-de-1996.pat).
I think the pattern you want is something like this:
.Da4tei5p4fad.
Here the odd numbers represent allowed hyphenation points, and the even numbers are disallowed hyphenation points.