Better Hyphenation

[roger64]

Originally Posted by JSWolf

.../...

The default hyphenation dictionary can be edited to give better results by having the numbers from 5 to 2.
.../...

Hi,

Thanks for this. Could you elaborate a bit on the above sentence?

I had prepared one for French language only (from Grammalecte/LibreOffice) last year which worked well for 2.8.1. and that I keep using.
https://www.mobileread.com/forums/sho...0&postcount=10

[JSWolf]

Originally Posted by JSWolf

The default hyphenation dictionary can be edited to give better results by having the numbers from 5 to 2.

The Adobe hyphenation diction as it comes has

LEFTHYPHENMIN 5
RIGHTHYPHENMIN 5

The configuration parameter lefthyphenmin specifies the minimal length of a word prefix that can be separated from a word through hyphenation. The righthyphenmin parameter specifies the minimal length for a suffix.

So you would need a word of at least 10 letters and it would only hyphenate if at least 5 letters were in front of the hyphen and at least 5 letters behind the hyphen. This basically can leave a lot of gaps in lines where a hyphen would normally work fine. I've used a left of 2 and a right of 3. I'm thinking of changing it to a right of 2. But I'll test it before posting an update.

[JSWolf]

I've updated the hyphenation dictionary in the first post. It's just US English. I;ve changed the settings to give even better hyphenation.

The original settings were...
LEFTHYPHENMIN 2
RIGHTHYPHENMIN 3
COMPOUNDLEFTHYPHENMIN 2
COMPOUNDRIGHTHYPHENMIN 3

I've changed the settings to be...
LEFTHYPHENMIN 2
RIGHTHYPHENMIN 2
COMPOUNDLEFTHYPHENMIN 2
COMPOUNDRIGHTHYPHENMIN 2

I have tested this and it works well. With a rather largish font size, it hyphenated the word happy as hap-py and if it had not, there would have been some gaps in the line.

So if you've installed the previous version, give the new version a go.

If anyone wants me to create a better hyphenation dictionary for another language, please find an OpenOffice compatible hyphenation dictionary, PM me a link, I'll edit the settings, make the TGZ file and post it in this thread.

[GeoffR]

There are a couple of issues to do with words that have trailing punctuation:

1. Both the epub and kepub reader seem to count the punctuation when deciding if the word is long enough to hyphenate. So for example with both HYPHENMIN parameters set at 5, everybody would not be hyphenated, but everybody, could be hyphenated as every-body,

2. The kepub reader can hyphenate in the wrong place if there is trailing punctuation, so with both HYPHENMIN parameters set at 5, everything, could be hyphenated as everyt-hing, by the kepub reader, but the epub reader would still hyphenate it correctly as every-thing,

There is not much that can be done about the second problem, but setting RIGHTHYPHENMIN one higher than LEFTHYPHENMIN can help a bit with the first.

(I use LEFTHYPHENMIN 3, RIGHTHYPHENMIN 4)

[JSWolf]

Originally Posted by GeoffR

There are a couple of issues to do with words that have trailing punctuation:

1. Both the epub and kepub reader seem to count the punctuation when deciding if the word is long enough to hyphenate. So for example with both HYPHENMIN parameters set at 5, everybody would not be hyphenated, but everybody, could be hyphenated as every-body,

2. The kepub reader can hyphenate in the wrong place if there is trailing punctuation, so with both HYPHENMIN parameters set at 5, everything, could be hyphenated as everyt-hing, by the kepub reader, but the epub reader would still hyphenate it correctly as every-thing,

There is not much that can be done about the second problem, but setting RIGHTHYPHENMIN one higher than LEFTHYPHENMIN can help a bit with the first.

(I use LEFTHYPHENMIN 3, RIGHTHYPHENMIN 4)

Thanks for that bit of information. If I see any issues with hyphenation because of the right set to 2, I'll change it back to 3. But for now, I'll leave it at 2.

[roger64]

Originally Posted by JSWolf

If anyone wants me to create a better hyphenation dictionary for another language, please find an OpenOffice compatible hyphenation dictionary, PM me a link, I'll edit the settings, make the TGZ file and post it in this thread.

Thanks both of you for your explanations and for the proposal. Here is one for the French language.

Sorry for the external link, for some reason, I can't upload to MR

https://mega.co.nz/#!5EECmTDQ!mD0aXJ...VVI1oFVLkjdYHE

[Teom@n]

Originally Posted by JSWolf

I don't think those dictionaries will work. Can you find one for LibreOffice or OpenOffice?

I have found a project for Libre Office plugin for Turkish. I installed it to Libre Office but I couldn't its folder.

I attached files for Open office, firefox and another.

[JSWolf]

Originally Posted by Teom@n

I have found a project for Libre Office plugin for Turkish. I installed it to Libre Office but I couldn't its folder.

I attached files for Open office, firefox and another.

The first file is a bunch of Java files (.jar) and no good. The second is just a dictionary without a hyphenation dictionary. Again, no good.

Have a look around, there might be a Turkish hyphenation dictionary for Tex and there is a way to convert a Tex hyphenation dictionary to one that is compatible with OpenOffice. Google is your friend.

[JSWolf]

First post updated with a French hyphenation dictionary.

[frostschutz]

Maybe this? http://code.metager.de/source/raw/mo...on/hyph_tr.dic

[JSWolf]

Originally Posted by frostschutz

Maybe this? http://code.metager.de/source/raw/mo...on/hyph_tr.dic

Maybe. Does anyone know what happens if an incorrect hyphenation dictionary is loaed on to a Kobo Reader? Does it lock or does it just ignore it?

[Teom@n]

Originally Posted by JSWolf

The first file is a bunch of Java files (.jar) and no good. The second is just a dictionary without a hyphenation dictionary. Again, no good.

Have a look around, there might be a Turkish hyphenation dictionary for Tex and there is a way to convert a Tex hyphenation dictionary to one that is compatible with OpenOffice. Google is your friend.

Ok, I am still searching. Google is my friend

[Teom@n]

Originally Posted by frostschutz

Maybe this? http://code.metager.de/source/raw/mo...on/hyph_tr.dic

maybe

[roger64]

Originally Posted by JSWolf

First post updated with a French hyphenation dictionary.

Jon

Thanks for your kind support!

[JSWolf]

Originally Posted by roger64

Jon

Thanks for your kind support!

Happy it worked. Enjoy!