Aura H2O Just bought a Kobo Aura H2O

[lvicini]

Originally Posted by davidfor

You can attach an epub here if it doesn't break copyright. If you stripped it to one page, you might be OK on "fair use" rules. But, you cold also replace the text with something else. And I feel I have seen a plugin or tool that randomises the text in an epub for just this sort of thing.

Hello here is the file. I stripped as much as content as possible. The problem is happening in the grey box. The <ul> (starting at line 19 within c00.xhtml) doesn't get displayed depending on the text size selected.

Let me know if you have any idea of what might be wrong.

[GeoffR]

That is an EPUB3 book, you shouldn't be too surprised if it doesn't work the same on an EPUB2 reader.

Edit: To clarify futher: iBooks and Kobo's KePub reader are (approximately) EPUB3 readers, whereas Kobo's ePub reader is EPUB2.

[davidfor]

Being epub3 is part of it, but I'm sure someone has said the version of the RMSDK that Kobo is now using has some epub3 support.

But, the real culprit is the following CSS:

.feature1,
.feature2,
.feature3,
.feature4 {
    border-radius: 8px;
    margin: 2px auto;
    padding: 8px;
    page-break-inside: avoid;
    text-align: center;
    width: 95%;
}

And specifically "page-break-inside: avoid". If I remove that, everything is OK.

I think that what is happening is that if the text with one of the above classes fits on a page, everything is OK. But, if it doesn't, it pushes the text to the next page where it doesn't so it all gets pushed to the next page and so on. The same thing happens if I change it to epub2. Definitely a bug, but removing the attribute will fix it.

Is this the book http://au.wiley.com/WileyCDA/WileyTi...19053064.html? And in the Kobo shop, https://store.kobobooks.com/en-us/eb...-integrations? I'll pass this along to Kobo, but I want to include the reference to the full book. And having it in the Kobo shop should allow them to get a copy for testing.

[Ken Maltby]

Looks ok, in Koreader. In both portrait and landscape. Most of the font sizes are working fine, but a few don't. This is with the embedded style on so that the grey box is displayed.

Luck;
Ken

[jackie_w]

Originally Posted by davidfor

And specifically "page-break-inside: avoid". If I remove that, everything is OK.

I vaguely remember reading elsewhere on these forums over the years that the Adobe renderer doesn't support "page-break-inside: avoid". If this is the case I'm not sure how much Kobo could do (or would be prepared to do) about it.

[lvicini]

Just few days ago, I was convinced that the EPUB format was based on PostScript... :-D

A couple of questions pop to my mind
1) Is the Kobo standard onboard EPUB reader KePub or ePub reader?
2) I read in Wikipedia that EPUB3 is based on HTML5. Considering that open source HTML5 web browser engines are everywhere, why are there still EPUB2 readers around? It should be easy to make all of them EPUB3 capable.

[frostschutz]

Originally Posted by lvicini

I was convinced that the EPUB format was based on PostScript... :-D

Unfortunately not the case. I liked PostScript, I wrote a Japanese manuscript paper generator in it... a long time ago

Originally Posted by lvicini

1) Is the Kobo standard onboard EPUB reader KePub or ePub reader?

It has both.

Originally Posted by lvicini

It should be easy to make all of them EPUB3 capable.

It's because of DRM. The EPUB reader software is by Adobe, each reader brand just builds their own user interface on top of it, but basically you can't separate the core functionality of Adobe DRM / EPUB / PDF support.

With Kobo you get EPUB3 support in a way, but it's KEPUB, their own format, their own DRM solution, ...

Now why Adobe did not add support for EPUB3 ages ago, who knows? Instead they created a new more secure DRM... thanks for nothing

[davidfor]

Originally Posted by jackie_w

I vaguely remember reading elsewhere on these forums over the years that the Adobe renderer doesn't support "page-break-inside: avoid". If this is the case I'm not sure how much Kobo could do (or would be prepared to do) about it.

That was my thought to, but it does seem to be doing something. And as to Kobo fixing it, no, I expect that they have to wait for it to be fixed in the RMSDK, so I don't expect it to happen any time soon.

[DNSB]

Originally Posted by lvicini

1) Is the Kobo standard onboard EPUB reader KePub or ePub reader?

Kobo uses two renderers. Adobe's RMSDK which handles epub2 and Adobe's ADEPT DRM. ACCESS's NetFront renderer which is more or less an epub3 compliant renderer.

Originally Posted by lvicini

2) I read in Wikipedia that EPUB3 is based on HTML5. Considering that open source HTML5 web browser engines are everywhere, why are there still EPUB2 readers around? It should be easy to make all of them EPUB3 capable.

Epub3 is based on HTML5. Unfortunately, I've tested several "epub3" renderers and the results were varied. Some came close to full compliance, other's were not much better than using Adobe's older epub2 only RMSDK to read an epub3 ebook.

You might want to check the support grid at epubtest.org, BISG EPUB 3 Support Grid. The best epub3 renderers are getting close but still not quite ready for prime time (IMNSHO) especially when you consider that these results are from test files that are epub3 compliant with epubcheck showing no errors. I've run the ~20 "epub3" ebooks I've purchased through epubcheck and none of them made it through error free. Some minor errors but ~33% had issues with the HTML navigation document (replaces the toc.ncx file in epub2 ebooks) while about 50% were epub2 ebooks with an epub3 navigation document.

Most of the errors in the navigation document were hard to understand since the standard is fairly easy to follow. Ordered list, optional header, li elements must contain a <a> or <span> with <a> elements capable of having ordered lists to allow for a contents hierarchy while the <span> elements must be followed by an ordered list. You can make a very complex TOC as in one book of poetry where the ~100 poems were grouped in a 3 level hierarchy or you can make a simple 1 level TOC.

As for the number of HTML5 renderers? Check the test results on those. Even the big names such as Google, Mozilla, Microsoft or Apple still have a ways to go to support HTML5 fully. The last range I saw was from 526 for Chrome 44, 525 for Opera 31 (perhaps using the same Chromium code base helped?), 478 for Firefox 44 (aka Nightly), 453 for Edge 13 and 336 for IE 11 -- out of a possible 555 points.

[lvicini]

Originally Posted by davidfor

Is this the book http://au.wiley.com/WileyCDA/WileyTi...19053064.html? And in the Kobo shop, https://store.kobobooks.com/en-us/eb...-integrations? I'll pass this along to Kobo, but I want to include the reference to the full book. And having it in the Kobo shop should allow them to get a copy for testing.

Yes it's the book. I bought it directly on the Wiley website. Not sure if the book has been corrected on the kobo shop