KF8 and hyphenation support

[pirl8]

As we all sadly know, Kindle does not jet suport hyphenation... Well, that's not completely true.

Actually the KF8/mobi7/AZW3 implemented in Kindle Fire and in Kindle touch since firmware 5.1.0 (and hopefully in K4 in the near future) has something that could be used as a workaround. Try the attached sample or look at the screenshots.

The 1st paragraph is hyphenated and hyphenation is dynamically rendered.

How? Simply inserting the SoftHYphen HTML entity: ­
It works very well (also note the DIV with rounded corners, which also are supported by KF8).

Actually this method is bulky and makes html code unreadeable but it works and doesn't create problems to the kindle searching algorithm.

I would like to implement a Calibre module to automatically insert SoftHYpens in generated AZW3 but I am completely new in Python programming and it may take me a long time (here is a thread with Kovid hints), so if someone feels to be able to do it, he's welcome.

Maybe this method could be used otherwise?

[JSWolf]

Oh heck no... this would seriously bloat things big time.

[tvjohn]

Originally Posted by pirl8

we all sadly know, Kindle does not jet suport hyphenation... Well, that's not completely true.

...(and hopefully in K4 in the near future) has something that could be used as a workaround. Try the attached sample or look at the screenshots.

The 1st paragraph is hyphenated and hyphenation is dynamically rendered.
..
Actually this method is bulky and makes html code unreadeable but it works and doesn't create problems to the kindle searching algorithm.

...

Works fine, including searching in K4NT 4.1.0

[pirl8]

Originally Posted by JSWolf

Oh heck no... this would seriously bloat things big time.

Why? Imagine it to be used wisely, for example just in text around pictures, or words longer than n characters. IMHO, it would give a better readability without bloating too much.

(Of course I'm talking in a Kindle user perspective)

[pdurrant]

Originally Posted by pirl8

How? Simply inserting the SoftHYphen HTML entity: ­
It works very well (also note the DIV with rounded corners, which also are supported by KF8).

I'll take occasional short lines over four hyphens in a row.

It would be much better if Amazon implemented a good hyphenation algorithm — a paragraph or perhaps page based one.

[pirl8]

Originally Posted by pdurrant

It would be much better if Amazon implemented a good hyphenation algorithm.

I do definitely agree. Even because "Kindle for PC", for example, actually already implements hyphenation, even on mobi6 files. So it seems to me very very silly that they choose not to officially implement it in KF8... Or maybe they had a very good reason, but I can't imagine which.

[pdurrant]

Originally Posted by pirl8

I do definitely agree. Even because "Kindle for PC", for example, actually already implements hyphenation, even on mobi6 files. So it seems to me very very silly that they choose not to officially implement it in KF8... Or maybe they had a very good reason, but I can't imagine which.

Just a clarification: The implementation usually happens in the renderer, not in the format.

Although it is possible to give a format specific ways to add hints for a hyphenation algorithm, in general it's not necessary. I'm not sure whether KF8 has any such hinting mechanism.

[pirl8]

Originally Posted by pdurrant

Just a clarification: The implementation usually happens in the renderer, not in the format

Yep. That's what I meant. AFAIK ePub has a CSS setting to tell the renderer to hyphenate or not. The hyphenation algorithm is implemented at some level in the renderer.

The fact that ­ is correctly rendered can be used as a workaround; for Amazon it would't be that hard to implement a simple solution: when parsing HTML (and I guess there's an XHTML parser somewhere) the ­ could be inserted in the right position (with language guessed from the book). The user could then decide to turn this feature on or off (exactly as we do with latest firmware for embedded fonts).

Maybe it's not a great implementation but I think that a page with hypens is more pleasant to read than one without, expecially when using enlarging the font.