Hyphenation under epub3

[Tex2002ans]

Originally Posted by JSWolf

I believe that search still works on a Kindle when soft hyphens are using with KF8. I do not know about the rest of the gotchas in your list.

Those glitches were a list of "soft hyphen" issues—across all sorts of software/programs—going way, way back to at least 2013/2015.

Soft hyphens really are an atrocious character that should be completely avoided.

And, within the past few years, Auto-Hyphenation in devices/browsers and word processors has gotten so much better—so if you're on a modern device or using recent apps/programs, you shouldn't have a need for manually inserting soft hyphens.

[JSWolf]

Originally Posted by Tex2002ans

Those glitches were a list of "soft hyphen" issues—across all sorts of software/programs—going way, way back to at least 2013/2015.

Soft hyphens really are an atrocious character that should be completely avoided.

And, within the past few years, Auto-Hyphenation in devices/browsers and word processors has gotten so much better—so if you're on a modern device or using recent apps/programs, you shouldn't have a need for manually inserting soft hyphens.

Well, you can buy the latest Kindle Paperwhite and KF8 still doesn't have hyphens. So if you read KF8, you either read without hyphens or your read with soft hyphens. I know KFX has hyphens, but KFX is an abomination.

[RbnJrg]

When I started this thread (in Feb-2019) I had asked how to enable hyphens under epub3 and in ADE 4.5x. Today I can answer that is possible to get that just with the following property:

p {
...
   -ms-hyphens: auto;
}

ADE 4.5x doesn't need more to enable hyphens with epub3. But now that this old thread has been resumed, I have another question related with hyphens and epub3. I agree with everything Tex2002ans said regarding soft hyphens. For that reason in the past year I was exploring the use of javascript to add hyphens by means of the script "hyphenator". On some ereaders, it works flawlessly (no issues with searching words or TTS) but on others (for example Calibre Viewer), it has serious drawbacks. Theorically, the issues of "hyphenator.js" were fixed by "hyphenoly.js"; I say "theorically" because I couldn't -so far- make it work in an epub3 (it works great in browsers).

By having said that, by chance, anyone here was able to implement "hyphenopoly.js" in an epub3?

[JSWolf]

It's a really bad idea to add in JavaScript for hyphenation. Either the program being used supports hyphens or it doesn't. I don't know of any program that supports soft hyphens as well as the KF8 renderer on a Kindle. So I would not use soft hyphens with ePub.

ADE 2.0.1 on up supports hyphenation. So no need for any code to turn on hyphens if using ADE. I don't know of any program that needs code to turn on hyphens other then maybe the KePub renderer (on a Kobo).

[Tex2002ans]

Originally Posted by RbnJrg

For that reason in the past year I was exploring the use of javascript to add hyphens by means of the script "hyphenator". On some ereaders, it works flawlessly (no issues with searching words or TTS) but on others (for example Calibre Viewer), it has serious drawbacks.

[...]

By having said that, by chance, anyone here was able to implement "hyphenopoly.js" in an epub3?

Hmmm. Not aware of any. You seem to be the one on MR testing and poking around with that bleeding-edge research.

What was the "serious" issues happening with Calibre's Viewer?

Originally Posted by JSWolf

It's a really bad idea to add in JavaScript for hyphenation. Either the program being used supports hyphens or it doesn't.

Yeah, in published books going for sale, I'd definitely err on that side too.

But RbnJrg works a lot on personal releases, so he can get away with more bleeding-edge EPUB3+Javascript work.

Side Note: I was never a big fan of these hyphenation shims. (Even on websites.)

And once I began messing around a lot more with Auto-Translation and features like that, I'm always leery on how those will interact with:

• Device/Browser hyphenation
• + Javascript hyphenation on top.

(Like let's say you have German text that you Auto-Translate to English. Will the (javascript) hyphenator still be applying German hyphenation to it?)

Lots of things that need to be poked around and tested.

Originally Posted by JSWolf

ADE 2.0.1 on up supports hyphenation. So no need for any code to turn on hyphens if using ADE. I don't know of any program that needs code to turn on hyphens other then maybe the KePub renderer (on a Kobo).

But that old ADE might support hyphenation in only a handful of languages. (Does anyone have an exact list of which languages are supported?)

Over the past few years, there's been lots of activity+updates in the Hyphenation Dictionaries.

Like within the past 2 years, there's been new/better support for:

• mr = Marathi
• fy = Frisian
• cs = Czech
• fi = Finnish
• mk = Macedonian

[JSWolf]

Originally Posted by Tex2002ans

But that old ADE might support hyphenation in only a handful of languages. (Does anyone have an exact list of which languages are supported?)

Over the past few years, there's been lots of activity+updates in the Hyphenation Dictionaries.

Like within the past 2 years, there's been new/better support for:

• mr = Marathi
• fy = Frisian
• cs = Czech
• fi = Finnish
• mk = Macedonian

ADE 2.0.1 supports English, German, Spanish, French, Italian, and Portuguese.

I don't know if I can add more languages by just adding in the hyphenation dictionary.

[RbnJrg]

Originally Posted by Tex2002ans

And once I began messing around a lot more with Auto-Translation and features like that, I'm always leery on how those will interact with:

• Device/Browser hyphenation
• + Javascript hyphenation on top.

(Like let's say you have German text that you Auto-Translate to English. Will the (javascript) hyphenator still be applying German hyphenation to it?)

When you add the Hyphenator script (that is deprecated), you also must add the pattern files that contains the rules to hyphenate the text. For example, if you want to hyphenate german text, besides the script, it must be included the file "de.js". If besides german text, you also want to hyphen english text, the file "en-us.js" also must be included. You can include all the pattern files you wish. German text gains a lot with hyphens. For example, some time ago, Doitsu uploaded a german epub3 called "Čapek, Karel: Der gestohlene Kaktus"; that epub has all the text aligned to left. Well, I took that epub and I added "hyphenator.js" and "de.js" so hyphens where enabled. You can see how better looks the things (and also to watch how easy is to add hyphens with javascript).

Čapek,_Karel-Der_gestohlene_Kaktus.epub

What was the "serious" issues happening with Calibre's Viewer?

Judge by yourself. Open the epub I attached in Calibre Viewer and try to pass all the pages one by one.

The problem with Hyphenator.js is that the script adds soft hyphens like "Hyphenate this" and so, it has the same issues with searching and TTS. On the other hand, Hyphenopoly.js adds an "aria-label" that would avoid those drawbacks.

Regrettably, not many programs and apps have a built-in hyphenator. For example, they can't hyphenate by themself THORIUM (!), Koodo (another great ereader for both Windows and Linux), Reasily, Lithium, Aldiko Next, Overdrive, Google Play, BookFusion, Infinity Reader and and the list goes on (the only ereader so far I found perfect was Foliate, for Linux). On Aldiko Next, for example, Hyphenator works great (no issue with searching and TTS) but on Reasily, it has issues with searching. Hyphenopoly theorically would solve the problems, but i cant get it to work on epubs because it doesn't recognize the user agent (maybe if I can fake the user agent, I can get something). Hyphenoly first tries to see if there is native support for hyphens; if so, then does nothing; on the contrary, adds soft hyphens with an aria-label to avoid issues with searching and TTS.

[Hitch]

Originally Posted by Tex2002ans

For CSS hyphens to work, it requires:

• Your device to have the proper Hyphenation Dictionaries for your language.
• Your ebook to be marked with the proper HTML language.

For example, Kobo has great support, and even allows you to add your own custom dictionaries for any language.

Kindles are locked down, so if Amazon hasn't added a hyphenation dictionary for (RandomLanguageX), you're out of luck.

- - - -

Side Note: According to jhowell in 2021, your Kindle should already support Italian hyphenation:



So, having no hyphenation automatically happening on your Kindle, something else must've been wrong.

- - -



I'd put a very huge WARNING label on this.

Only use Hyphenate This! on an ebook:

• You are going to be personally reading.
• As a very, very last resort.
  
  • Like if your ereader is really old.

DO NOT ever use it on a book you're trying to sell.

It completely fills the book full of SOFT HYPHENS... and while it may LOOK like "hyphenation now works", it can cause many other problems with:

• Search
• Spellchecking
• Copy/Pasting
• Dictionary popups
  
  • Where you press+hold on words.
    
• [...]

(More disadvantages listed in my 2021 post.)

It's much better to do hyphenation automatically + on your actual device.

(Luckily, things have been getting much better in the past few years!)

- - -

Side Note #2: For all the hyphenation info you'll ever need to know... I've written A TON about it:

• 2021: "Welsh"
• 2021: "Add xml:lang to ePub"
• 2021: "Does hyphenation work or not?"

I'd also recommend just typing this into your favorite search engine:

hyphenation Tex2002ans site:mobileread.com

The past few years, I've been really explaining why:

• Marking your ebook + HTML language properly

is super important. And proper hyphenation is one of the key reasons!

Using Send to Kindle won't work, typically, from the jump, because it does not support Enhanced Typesetting. ET is not supported, as it happens, KDP system/eco-system wide. They explain this in no uncertain terms during the file export options from KP3.

The only reason that any hyphenation is showing up, at all, is due to "Hyphenate this!" being used, via Calibre. It has no relationship to, and won't resemble, what will really happen when the ePUB is uploaded to KDP and rendered into the latest and greatest format, WITH ET.

That's the short of it.

Hitch