From ePUBSecrets: The state of ePub standards support

[fjtorres]

While discussing myths about ePub, ePUBSecrets presents a pretty concise update on how thoroughly the major vendors support ePub specs and how much cross-vendor compatibility exists today:
http://epubsecrets.com/reviewing-epu...p#comment-3288

On August 22, Aptara (@aptara) hosted a Webinar titled “eBook & App Device Compatibility: Separating Fact from Fiction” with Joshua Tallent (@jtallent), CEO and Founder of eBook Architects, and Jean Kaplansky (@JeanKaplansky), a Solutions Architect at Aptara. It was an informative Webinar, and if you missed it you can view the recording by registering here: https://event.on24.com/eventRegistra...epage=register

You should also be able to download the slides from the presentation once you launch the Webinar. I encourage you to watch the Webinar since there is a lot of great content in it on the current state of eBooks and what is and is not supported.

Tallent and Kaplansky looked at reflowable files and fixed-layout files on the major retailers, with Tallent covering Kindle and iBooks and Kaplansky covering B&N’s Nook and Kobo. The Webinar used a Myth and Reality format to share the information. Here are the myths that two of them presented.

The whole article is worth reading. For some, tracking down the webinar may be worth the while.

Conclusions?

Tallent’s key takeaways included:
Reflowable files are the best option for the majority of eBooks.
Kindle Format 8 (KF8) is NOT EPUB.
Fixed layout: children’s titles work well in Apple, Amazon, B&N; non-fiction fixed layout currently only works well in Apple.

Kaplansky’s key takeaways included the following:
There will never be “one EPUB file to rule them all.” Retailers and device manufacturers differentiate themselves and their product by the details of their implementations. Be aware of the details!
Handle fixed layout with care and planning. Know your target devices and do what is right for the text rather than the design‑unless the two are intrinsically linked.
Users prefer annotation and highlighting functionality over fixed layout when it comes to textbooks and non-fiction.

The Upshot

So what does it all mean? These were my takeaways on the myths and Tallent and Kaplansky’s takeaways.

1. The only retailers that matter are Amazon, Apple, B&N, and Kobo

While the Webinar looked at fixed layout and reflowable and what each retailers supports, and even more telling item is the four retailers that were selected. There were other out there, but these are the four for whom eBook designers are currently designing. They are setting the agenda, and ultimately, they will determine if EPUB3 in its entirety ever comes to maturity in the marketplace.

2. We are currently caught in a limbo state between EPUB 2.0.1 and EPUB3

3. The major retailers are moving further apart, not coming together over EPUB3

4. The major retailers do not have the same support across their own devices and apps

The major retailers have multiple devices with different capabilities in the
marketplace and have apps for all the major platforms. The retailers support of their own implementations is not 100 percent across all of their devices.
Providing a single EPUB to each vendor even with media queries to determine what kind of device the reader has is difficult at best.

5. Outside of iBooks Author, fixed layout for non-fiction is probably not a good idea


6. This does not scale and removes the cost benefits of EPUB

Publishers can not afford to produce four different versions of EPUB for four different retailers (and more for additional retailers in the future). The cost of EPUB creation will increase as the retailers specs diverge further. This will likely mean many pubs will make a choice of targeting Amazon and KF8, and a single other retailer, likely Apple, with a single EPUB file. The other retailers will have to take what they can get in terms of the EPUB or go without. The future this leads to: Amazon and Apple will be the main, possibly only, eBook retailers.

7. Publishers and the IDPF need to pressure retailers to support and accept the full EPUB3 spec now or it will never be implemented

We are at a crossroads. Either EPUB3 is supported in the marketplace and the full spec is implemented by retailers, allowing publishers to produce one file for everyone, or the retailers will continue to take a pick-and-choose approach to the spec, leading to different formats and different capabilities for each retailer’s devices and apps. It is time for publishers and the IDPF to exert whatever influence they have to right the EPUB ship, convince the retailers to fall in line behind the full EPUB3 spec, and prevent a further splintering of the format.

My takeaways are shorter:
1- ePub is far from being a universally-compatible format for commercial ebooks even in the non-Kindle market.
2- The lack of teeth behind the IDPF approach to "standards" is coming back to haunt everybody.
3- For the forseeable future, talk of "ePub" at the consumer level will mean ePub 2.0.1 not ePub3.

I'm not ready to say that long-term the market for commercial ebooks will come down to just Amazon and Apple but that outcome is not out of the question. Especially if Apple prices the mythical iPad Mini low enough and locks out other reader apps from its appstore.

Comments?
Alternate data sources?

[Dave_S]

After reading all that, I am glad that I gave up on the EPUB non-standard and archived all my books in FB2 format. If EPUB ever does become a real standard then it should be easy to convert to it then.

[DrNefario]

EPUB is based on HTML and CSS, and there have never been any problems with the implementations of that.

That's sarcasm, by the way.

[fjtorres]

Originally Posted by DrNefario

EPUB is based on HTML and CSS, and there have never been any problems with the implementations of that.

That's sarcasm, by the way.

No sirree, bob.
Never, ever.
And embrace-and-extend isn't a standard business practice in the tech industries.
That is why tech standards *never* come with strong licensing enforcement.
(Hey, sarcasm is a fun way to kill time on mondays!)

All things considered, by the time the dust settles in the next decade, we may end up with SGML as the true universal standard.

[stewacide]

Given that Epub including Epub3 is just a subset of the web standards every modern browser supports, shouldn't it be absolutely trivial for Amazon/BN/Kobo/Sony to simply port Webkit or Mozilla to their devices, with some tweaks, as the rendering engine, a la Apple's solution with iBooks? Writing an Ebook engine from the ground up seems like re-inventing the wheel.

[tubemonkey]

My ePub problem was solved when I settled on Kindle.

[teh603]

Originally Posted by DrNefario

EPUB is based on HTML and CSS, and there have never been any problems with the implementations of that.

That's sarcasm, by the way.

Yeah... I remember those days. I also remember a few tools made to purge CSS code from an HTML document so it would actually work.

I may end up going with tubemonkey and doing my next book for the Kindle first, before Smashwords and BookieJar. That means I'll have to get one, though. /sweatdrop

[JSWolf]

Originally Posted by teh603

Yeah... I remember those days. I also remember a few tools made to purge CSS code from an HTML document so it would actually work.

I may end up going with tubemonkey and doing my next book for the Kindle first, before Smashwords and BookieJar. That means I'll have to get one, though. /sweatdrop

You don't publish direct to B&N and Kobo?

[teh603]

Originally Posted by JSWolf

You don't publish direct to B&N and Kobo?

Right now I go thru Smashwords.

[fjtorres]

Originally Posted by stewacide

Given that Epub including Epub3 is just a subset of the web standards every modern browser supports, shouldn't it be absolutely trivial for Amazon/BN/Kobo/Sony to simply port Webkit or Mozilla to their devices, with some tweaks, as the rendering engine, a la Apple's solution with iBooks? Writing an Ebook engine from the ground up seems like re-inventing the wheel.

Not trivial.
There is simply too much room for interpretation of the specs that results in *visibly* discordant formatting in different engines/apps. That is why they talk of having to do different versions for all four of the retailers (and probably a fifth for generic Adobe clients if you expand their analysis outside NorthAm).
More, some of the epub features in epub3 (the primary subject of the forking explosion) are *very* CPU/GPU/DRAM intensive. There is stuff in epub3 that*Apple* is ignoring in their ibook fork that would eat alive current top-of-the-line iPads.

[teh603]

Originally Posted by fjtorres

Not trivial.
There is simply too much room for interpretation of the specs that results in *visibly* discordant formatting in different engines/apps. That is why they talk of having to do different versions for all four of the retailers (and probably a fifth for generic Adobe clients if you expand their analysis outside NorthAm).
More, some of the epub features in epub3 (the primary subject of the forking explosion) are *very* CPU/GPU/DRAM intensive. There is stuff in epub3 that*Apple* is ignoring in their ibook fork that would eat alive current top-of-the-line iPads.

Is it possible that people are asking too much of epub3 by setting a standard that would overwork a tablet?

Then again, it would be fairly par for the course. People like new stuff that pushes existing hardware to the breaking point, just because they can. Remember when Quake 4 won the Best Slideshow Game award?

[fjtorres]

Originally Posted by teh603

Is it possible that people are asking too much of epub3 by setting a standard that would overwork a tablet?

Yes, it is possible.
It is also possible that epub3 has been so future-proofed and over-spec'ed by trying to meet every conceivablepublisher requirements that it has overshot the needs of the current marketplace.

I've noticed a tendency over the past few iterations to try to turn epub into reflowable pdf, by asking it to support sophisticated fixed formatting as well as sophisticated smart reflow. And then they added interactivity, "spyware"-enabling programability, multimedia, etc.
That is a lot to ask of what is, as pointed out, a web-based spec built off HTML, CSS, and javascript.

ePub3 is not the first document standard that floundered for trying to be all things to all people--hence my sarcastic quip about SGML above. The reason MS DOC and TeX, first, and later HTML/XML, became the defacto industry standards that SGML aspired to be is because SGML overtaxed 80's-90's desktop PCs with its attempt to map all types of documents, from spreadsheets to scientific reports. Most business chose to go with much cheaper PCs and mainstream Office productivity apps instead of the expensive UNIX Workstations and desktop-publishing software needed to work with SGML.

So far, ebook vendors are (apparently) choosing to implement strategic subsets of epub3 functionality that their hardware devices can support and that meet their needs as commercial ebook vendors. Thus, Apple's ibook format supports the features they need for textbooks and Amazon supports the features *they* need for childrens' books and graphic novels.

That is exactly what happened to SGML; subsets of the full over-reach Spec were carved out for separate markets and HTML flourished in the web space. Eventually XML emerged out of market needs.

It may be that the committee needs to retrench and spec out an ePub 2.5 that can run on the existing installed base or readers and tablets.

Or maybe the ebook vendors are just greedy blood-suckers out to destroy a perfectly viable universal standard out of evil self-interest.

[StoryEnthusiast]

Originally Posted by fjtorres

All things considered, by the time the dust settles in the next decade, we may end up with SGML as the true universal standard.

Wow, when was the last time I read about SGML? Before the rise of HTML. Can we really go backward?

[fjtorres]

Originally Posted by StoryEnthusiast

Wow, when was the last time I read about SGML? Before the rise of HTML.

I'm dating myself, huh?
Or maybe I just know a good historian?

Originally Posted by StoryEnthusiast

Can we really go backward?

It's been known to happen.
Or...

Maybe the industry is just "doomed to repeat the errors of the past".
That dream of "one format to rule them all"--emphasis on rule--just doesn't go away.

[teh603]

Originally Posted by fjtorres

Or maybe the ebook vendors are just greedy blood-suckers out to destroy a perfectly viable universal standard out of evil self-interest.

Just this once I'm going to reverse my usual stance, and say that the standard wasn't viable to begin with. The vendors are simply trying to get something usable out of it.