Better Hyphenation

[Pioneer129]

Originally Posted by JSWolf

I've updated the English hyphenation dictionary in the first post of this thread. Please download it and install it. It should work better with KePub and work just as well with ePub.

Thank I have downloaded it and wil give it a try.
Have you any idea about dictionary lookup history on kobo nickle?

[JSWolf]

Originally Posted by Pioneer129

Thank I have downloaded it and wil give it a try.
Have you any idea about dictionary lookup history on kobo nickle?

I have no idea about the dictionary lookup history.

[Pioneer129]

Originally Posted by JSWolf

I have no idea about the dictionary lookup history.

I don't wanna bother you but might I ask another question?
Yesterday, I generated koboRoot Patch file after a lengthy process (for the first time) and then copied it into .kobo directory of my device. I have two questions in this regard.
1. The koboRoot Patch file is now not to be seen where I had copied it. Can you tell me what happened to it?
2. In future, if I want to add some new patches to my kobo, would I be doing the whole process right from the start (downloading patches and firmware and then generating koboRoot patch file) or is there some other way that I can do it?
e.g I want to enable the patch "disable update via wifi", how will I do it? Do I have to generate another koboRoot patch file for it and then copy it to .kobo directory or there is another way for it?

Thanks in anticipation

[MGlitch]

1) the system on the kobo unpacked the file and distributed the contents appropriately based on how they were stored
2) yes you’ll have to do most of that work again. There’s a file in the base kobo patch folder that overrides the “no”s from the files in the src folder but I’ve never seen much use for this myself, partially because I modify some of the patches to my tastes anyway and thus need to go into those files and also because I’ve not taken the time to really learn how this file is used.

But this really isn’t the thread to talk about this in. There’s the patching instructions thread which would be more relevant and more likely to get you full answers

[tartanpion]

Some news. First in contrary as I have said for the no break space bug : no solution with the "libroffice" hyphenFRdict old or beta.(hyphen respectively 2 and 3)
The "adobe" kobo hyphendict _FR is : right 3 left 3 and not 5 like the US dict.
All my tests were with 2 or 3 from the beginning.
Finally, I have seen this post :

Originally Posted by GeoffR

That was fixed in firmware 3.15.0.


Edit: When changing the LEFTHYPHENMIN and RIGHTHYPHENMIN values in the hyphenation dictionary, beware that the KePub and ePub readers both use the same dictionary but in recent firmware they interpret these parameters differently -- the KePub reader counts the value as one less than specified, so if you read KePubs then make the values one higher than the value you actually want. If you read both KePubs and ePubs then you'll need to find a compromise.

(And personally, I find full justification and hyphenation to be absolutely essential features. I'd go back to reading paper books rather than read ebooks left-justified or unhyphenated.)

I made my own dict with (and coumpound/compound corrected)

LEFTHYPHENMIN 4
RIGHTHYPHENMIN 3
COMPOUNDLEFTHYPHENMIN 4
COMPOUNDRIGHTHYPHENMIN 3
%%%   Do not remove the following line   %%%
NEXTLEVEL

test seems ok but I need more testing
and I'll test some extra css

body {-ms-hyphenate-limit-lines: 3; -webkit-hyphenate-limit-lines: 3; hyphenate-limit-lines: 3;}

And the hyphendict can have a nohyphen rule https://code.woboq.org/libreoffice/l.../hyphen.h.html but it's out of my skills...

[Leonatus]

@JSWolf: Jon, you have provided me with your tool for german language, too, and I thank you for it. However, there is an issue that is strange: When opening a german book (not so much in a french book that I'm reading simultaneously), it takes a considerable time until the page shows. That has not been the case before.

And - but this is supposingly not the issue of the tool - there occur strange hyphenations in the german books, for example: gram-m, or "sch" becomes hyphenated, what is, probably in 99,9 % of all cases illicit, for example: sc-h. I would like to edit the hyphenation dictionary, such as tartanpion has just reported, but could, please, someone tell me how to do?

[JSWolf]

Added a German hyphenation dictionary. See the first post in this thread.

[Krios]

Hi, i want to make english hyphenation dictionary for myself, right now i have hyphenEN_US.dic, hyphenEN_GB.dic and hyphenEN_CA.dic, so should i delete duplicate entries in them and combine them into one file, or should i just chose one and use that.

[JSWolf]

Originally Posted by Krios

Hi, i want to make english hyphenation dictionary for myself, right now i have hyphenEN_US.dic, hyphenEN_GB.dic and hyphenEN_CA.dic, so should i delete duplicate entries in them and combine them into one file, or should i just chose one and use that.

You could try combining them into one and deleting the duplicates and see how you get on. But add the following to the top of the dictionary...

UTF-8
LEFTHYPHENMIN 2
RIGHTHYPHENMIN 2
COMPOUNDLEFTHYPHENMIN 2
COMPOUNDRIGHTHYPHENMIN 2

[Krios]

Originally Posted by JSWolf

You could try combining them into one and deleting the duplicates and see how you get on. But add the following to the top of the dictionary...

UTF-8
LEFTHYPHENMIN 2
RIGHTHYPHENMIN 2
COMPOUNDLEFTHYPHENMIN 2
COMPOUNDRIGHTHYPHENMIN 2

Thanks, it had some problems at first, it would hyphenate, we'll and you've just before apostrophe, but i managed to fix, and now it works great.

[Buhaj47]

EDIT: I got it sorted out. Does re-uploading the archive file to the root folder replace the previously used dictionary?

If I add a Polish hyphenation dictionary, will it add the language to the device or replace one of the existing hyphenation dictionaries?

I have an attached dictionary file that says UTF-8 - is it okay to use it or should I pehaps use an older version with the ISO code?

[Moonbase59]

I quite like the Tolino and Kobo range of products but currently don’t have a Kobo at hand.

Since I just created complete new German hyphenation patterns and dictionaries for the Tolinos, addressing some of the problems @Leonatus mentions, I wonder if someone could test if those might also work on the Kobos.

Apart from the Hunspell- (Hyphen-)compatible formats (hyph_de.dic), Tolinos also seem to use the Hyphenator Javascript (de.js pattern file) which I also created.

Please refer to my (German-language) post Tolino - Deutsche Silbentrennung, an archive containing both files is attached there. It uses 3/3 hyphenation, both to address some problems with attached punctuation and to reduce the number of "2-letter hyphenations" which can slow down the reading flow.

If this works out for use on Kobo devices, I’d be happy if you’d include it in your installation software. The files contain English-language comments, and are based on TeX’s dehyph-exptl hyphenation (a 500,000+ word corpus), with additions to the word list, and recompiled patterns plus appropriate conversion to Hyphen and Hyphenator formats, with manual changes for Tolino quirks.

[Buhaj47]

What entry should I add to a dictionary to hyphenate maximum 2 consecutive lines? Here's the beginning of my file:

UTF-8
LEFTHYPHENMIN 3
RIGHTHYPHENMIN 3
COMPOUNDLEFTHYPHENMIN 3
COMPOUNDRIGHTHYPHENMIN 3
.a1
.a2b1s

Majority (but not all) entries in the dictionary start with a dot, what is it for?

[Semwize]

Originally Posted by Buhaj47

what is it for?

https://theswissbay.ch/pdf/Gentoomen...he_TeXbook.pdf

page 448

[Moonbase59]

Originally Posted by Buhaj47

What entry should I add to a dictionary to hyphenate maximum 2 consecutive lines? Here's the beginning of my file:

UTF-8
LEFTHYPHENMIN 3
RIGHTHYPHENMIN 3
COMPOUNDLEFTHYPHENMIN 3
COMPOUNDRIGHTHYPHENMIN 3
.a1
.a2b1s

Majority (but not all) entries in the dictionary start with a dot, what is it for?

You can’t (correctly). The whole dictionary would have to be recompiled in order to make a functional whole. The patterns (a leading dot means "at word begin") are calculated from a large word list, and several exception, prefix, suffix and rules lists), and later recalculated to follow a scheme the hyphenation file format creator invented (apparently for better performance).

Plus, some reader engines (older RMSDK, like in the Tolino) don’t recognize word boundaries correctly, resulting in bad hyphenation, because these engines "think" a punctuation character (like .,;»«›‹„“”‚‘’–—) directly adjacent to a word belongs to that word, which results in horribly bad "hyphenation" like "P-ferd" or "wil-l." (German words for "horse" and "want").

EDIT: Forget all that—I misread "consecutive lines" and somehow assumed "2-character" hyphenation. Lines you’ll specify in the book’s CSS (if EPUB), if your reader honours it, like so (for a paragraph):

p {
  /* hyphenation on */
  -webkit-hyphens: auto;
  -moz-hyphens: auto;
  hyphens: auto;
  /* hyphenation for max. so many consecutive lines */
  -webkit-hyphenate-lines: 2;
  hyphenate-lines: 2;
}