KOReader and epub3

[RbnJrg]

Please, add support for epub3 and css3 properties; also for javascript. If that can be done, KOReader will be the best reader for epub3 by several bodies of advantage.

[kodermike]

Originally Posted by RbnJrg

Please, add support for epub3 and css3 properties; also for javascript. If that can be done, KOReader will be the best reader for epub3 by several bodies of advantage.

Someone already asked via issue. Most recent response: https://github.com/koreader/koreader/issues/13115

[DNSB]

For what it may be worth, I would like to see ePub3 support being baked into KOReader. It's not as if the ePub3 specification is new since the first version (3.0) was finalized in Oct., 2011.

Thorium does a decent job of supporting ePub3 on my desktop and Kobo's proprietary renderer (the one used for kepub files on firmware 4 or less and used for kepub and non-drmed epubs on firmware 5.10+) does almost as decent job on my Kobo ereaders.

[RbnJrg]

Originally Posted by kodermike

Someone already asked via issue. Most recent response: https://github.com/koreader/koreader/issues/13115

What a shame! Evidently, the author has a misconception about what epub3 is, since he associates it only with a fixed-layout epub, and nothing could be further from the truth. Attached to this comment is a sample of my epub3 version of "Alice in Wonderland"; anyone who opens it on an epub3-compatible ereader (Calibre Viewer, any of the Sigil plugins, Thorium, Kobo for Android, etc.) will see that it's a fluid-layout epub, displaying text and images in a way that KOReader can't currently do, and that the layout is maintained regardless of the user changing the font size.
When KOReader can prevent fragmentation in an ebook like an epub3-compatible reader can, then it will be one of the best ereaders. But it doesn't seem like that's going to happen. What a shame.

[Frenzie]

Originally Posted by RbnJrg

What a shame! Evidently, the author has a misconception about what epub3 is, since he associates it only with a fixed-layout epub, and nothing could be further from the truth.

The discussion is about the cover page on the " 'advanced epub' from standardebooks.org", not about EPUB3. But the underlying point is the same of course: you need a full browser engine for it, and you're not going to get a full browser engine at this time.

Attached to this comment is a sample of my epub3 version of "Alice in Wonderland"; anyone who opens it on an epub3-compatible ereader (Calibre Viewer, any of the Sigil plugins, Thorium, Kobo for Android, etc.) will see that it's a fluid-layout epub, displaying text and images in a way that KOReader can't currently do, and that the layout is maintained regardless of the user changing the font size.

That's a bit lacking in specificity. You're likely to gain more by asking for specific properties or syntax, if only in the form of a more detailed explanation about how it would be hard to implement.

[Jellby]

Originally Posted by RbnJrg

anyone who opens it on an epub3-compatible ereader (Calibre Viewer, any of the Sigil plugins, Thorium, Kobo for Android, etc.) will see that it's a fluid-layout epub, displaying text and images in a way that KOReader can't currently do, and that the layout is maintained regardless of the user changing the font size.

Is that because of epub3 or because of CSS 3? Supporting an epub3 structure doesn't immediately mean supporting all fancy CSS 3 properties (or javascript). Support for CSS 3 can be added even to an epub2 reader. The problem is (I guess) that it's far from trivial, which is why a browser engine is usually employed.

I'd start with the "basics", and ask for vh, vw, flex and grid support, that should keep many people busy for long enough

[RbnJrg]

Originally Posted by Frenzie

That's a bit lacking in specificity. You're likely to gain more by asking for specific properties or syntax, if only in the form of a more detailed explanation about how it would be hard to implement.

The sample is not protected, so anyone can read the source. It's easy to see that KOReader can't support the "float" property, like the <figure> tag needs (the way the program handles this property so far generates blank spaces, which is a major flaw).

[RbnJrg]

Originally Posted by Jellby

I'd start with the "basics", and ask for vh, vw, flex and grid support, that should keep many people busy for long enough

I'd start with a better support for "float". And support for flex-box. To start, it would be enough for me.