EPUB 2 vs. 3 ??

[dan2the6th]

I've been seeing notes on new features in calibre to convert from EPUB v2 to v3. Which leads me to ask--when I convert a Kindle book to EPUB, which version is created? And why would I want to convert a book from EPUB v2 to v3?

[jackie_w]

Originally Posted by dan2the6th

I've been seeing notes on new features in calibre to convert from EPUB v2 to v3. Which leads me to ask--when I convert a Kindle book to EPUB, which version is created?

EPUB2

Originally Posted by dan2the6th

And why would I want to convert a book from EPUB v2 to v3?

I'd say it's one of those self-answering questions. If you don't know why you need EPUB3, you probably don't need it.

[kovidgoyal]

The only people that really need EPUB 3 are people that want to create EPUBs that have to pass epubcheck. Passing epubcheck is meaningless, but unfortunately some EPUB stores require it so people that want to sell books in those stores need it.

Other than that pretty much everything you can do in EPUB 3 you can do in EPUB 2, because both are ultimately rendered using the same browser engines.

[DNSB]

Originally Posted by kovidgoyal

The only people that really need EPUB 3 are people that want to create EPUBs that have to pass epubcheck. Passing epubcheck is meaningless, but unfortunately some EPUB stores require it so people that want to sell books in those stores need it.

Other than that pretty much everything you can do in EPUB 3 you can do in EPUB 2, because both are ultimately rendered using the same browser engines.

Minor nits. Epubcheck will pass an epub2 as easily as an epub3. Both versions are supported -- somebody had to use the version information.

As for rendering epub2 vs. epub3, there may be a renderer that handles both versions properly but I haven't found it yet. Likely the reason that Kobo uses two renderers with the software for their eInk devices.

The only epub3 books I've found that really needed to be epub3 are either displaying languages that don't read RTL, top to bottom (Japanese vertical with furigana as an example) or are using fixed layout.

[kovidgoyal]

Originally Posted by DNSB

Minor nits. Epubcheck will pass an epub2 as easily as an epub3. Both versions are supported -- somebody had to use the version information.

The point is that you can use EPUB 3 features in EPUB 2 books (the HTML/CSS/javascript features in particular) and they will work everywhere that EPUB 3 will work, except that epubcheck will complain about them.

As for rendering epub2 vs. epub3, there may be a renderer that handles both versions properly but I haven't found it yet. Likely the reason that Kobo uses two renderers with the software for their eInk devices.

Really, there are epub 3 renderers that do not render epub 2? If so, please point me to one. IIRC the reason Kobo does that has nothing to do with epub 3 they split their renderers long before epub 3, to support kepub.

[DNSB]

Originally Posted by kovidgoyal

The point is that you can use EPUB 3 features in EPUB 2 books (the HTML/CSS/javascript features in particular) and they will work everywhere that EPUB 3 will work, except that epubcheck will complain about them.

As you said, using epub3 features in epub2. You are going to run into epub2 renderers that will not behave properly when those features are used -- failing gracefully is not in their design.

Originally Posted by kovidgoyal

Really, there are epub 3 renderers that do not render epub 2? If so, please point me to one. IIRC the reason Kobo does that has nothing to do with epub 3 they split their renderers long before epub 3, to support kepub.

Kobo selected the ACCESS NetFront™ BookReader v1.0 EPUB Edition (in future, referred to as ACCESS) as an epub3 renderer. To quote part of a press release from 2012: "preparing to launch the Kobo eReading solution in Japan by implementing the NetFront BookReader v1.0 EPUB Edition, as a standard viewer engine for EPUB 3-compatible services, which can accurately reproduce complex Japanese characters and page layouts."

As for epub3 renderers that don't support epub2? The question for me is not if they support epub2 but if they support it to the point where I can't tell if I have an epub2 or an epub3 ebook open in calibre's ebook viewer, Datalogics DLReader, Edge, ADE4, Azardi, Readium, BlueFire and ghod alone knows how many others I've tried on my Windows desktop or in RMSDK and ACCESS on my Kobo ereaders. Some come very close but there are items that don't render the same and I find those annoying much as I find applications that do not follow the epub stylesheet(s) to be annoying. But then I also find Adobe's inability to support small caps as annoying.

Every now and then, I get enthused and try the epub3 test suite just to see how much has changed.

Last time through, I added "Harry Potter: A Journey Through a History of Magic" from the British Library as a real word fixed layout epub3 (aka pdf for epub users). Looks great on my Kobo after being renamed to .fxl.kepub.epub (uses the ACCESS renderer in FLO mode), Edge opened it in a two page layout which looked good. Readium for Chrome insisted on unpacking the file but did a nice 1 page view. Most of the other applications I tested strove to look ugly at best.

[davidfor]

Originally Posted by DNSB

Kobo selected the ACCESS NetFront™ BookReader v1.0 EPUB Edition (in future, referred to as ACCESS) as an epub3 renderer. To quote part of a press release from 2012: "preparing to launch the Kobo eReading solution in Japan by implementing the NetFront BookReader v1.0 EPUB Edition, as a standard viewer engine for EPUB 3-compatible services, which can accurately reproduce complex Japanese characters and page layouts."

KePub existed a long time before that. I've always taken them as for Kobo not to give money to Adobe. Books downloaded through Kobo devices and apps shouldn't suffer the license fees. Plus, I think it is a way for them to keep control and offer more features. There seems to be more limitations involved with the RMSDK than ACCESS.

[DNSB]

Originally Posted by davidfor

KePub existed a long time before that. I've always taken them as for Kobo not to give money to Adobe. Books downloaded through Kobo devices and apps shouldn't suffer the license fees. Plus, I think it is a way for them to keep control and offer more features. There seems to be more limitations involved with the RMSDK than ACCESS.

Before adding the ACCESS renderer, I seem to remember RMSDK was used for both epub and kepub with the kepubs being decrypted before sending to RMSDK.

[dan2the6th]

Thanks to all for the helpful info!