View Full Version : Why is the EPUB format so poorly implemented (on the desktop at least) ?


DSpider
02-14-2011, 05:25 PM
I mean come on. It's HTML with a little bit of CSS.

YfSNoLZKSio

Half a year later I expected a lot of desktop readers to show up. Multi platform readers. Perfect readers. Because what looks ok in one reader, doesn't look so well in another. Why is it so hard to develop a piece of software that displays the same like most web pages do today ?

Even the readers themselves look [vulgarity deleted - MODERATOR]. FBReader, EPUBReader (Firefox Add-on), even Adobe Digital Editions !

I expected EPUBReader to work perfectly since Firefox is such a well developed browser. Nope.

I also expected Adobe Digital Editions to work perfectly since I heard the "engine" is implemented in most e-Ink readers and they're the only ones with nasty DRM schemes for EPUBs. But no. They obviously focused more on the interface instead of functionality and ease of deployment. Deployment, yeah. It's probably the most most disgusting thing I ever installed on my computer. [vulgarity deleted - MODERATOR] is up with the install process ?? It installs through flash, through the [vulgarity deleted - MODERATOR] web browser ??? I was expecting a [vulgarity deleted - MODERATOR] .exe like the flash plugin but nooooo, if you're using Linux it doesn't give you a link or something but they [vulagarity deleted - MODERATOR] to say "only install from sources you trust" ? Like Adobe ? Gtfo...

The only desktop reader I kind of, sort of like is the "Kindle for PC". It's not perfect but (except for the icon and the very small do not register button when you first open it) you can tell it's well developed. Why isn't there a clone, multi platform or not, that supports EPUB ?

And why doesn't Amazon support EPUB ? What's their deal ? Are they afraid 5 years from now the dudes who conjured up EPUB will put a lock on it and ask Amazon to pay royalties or something ? They support HTML, don't they ? They do, yeah; they have a website.

wallcraft
02-14-2011, 07:59 PM
There are now a few implementations of mobile ADE on the Desktop, see Reading Adobe ePub on the Desktop (http://www.mobileread.com/forums/showthread.php?t=111863). The lack of Linux support from Adobe is notable, particularly since there are now several Linux based devices with mobile ADE and WiFi-based download capability - but no similar support for the Linux Desktop.

Another option for DRM-free ePub is Calibre's Reader.

Toxaris
02-15-2011, 04:10 AM
What do you mean by poorly implemented? Most readers do it reasonably well. I agree that there are not a lot of readers there, but then again, the format is more for portable devices like readers.

pdurrant
02-15-2011, 04:13 AM
I mean come on. It's HTML with a little bit of CSS.


This is also the main problem. It's XHTML 1.0 with CSS 2.0, but not quite. There are some bits it doesn't support, and some minor extensions.

So a proper ePub reader shouldn't just use a browser engine, but one specifically designed for the ePub dialect. Unfortunately, no-one does, and no-one supports everything that ePubs should support.

Jellby
02-15-2011, 05:02 AM
Because no one cares, they only want to include video and sound in books :rolleyes:

amoroso
02-15-2011, 05:47 AM
I mean come on. It's HTML with a little bit of CSS.
I was expecting a god damn .exe like the flash plugin but nooooo, if you're using Linux it doesn't give you a link or something but they have the balls to say "only install from sources you trust" ?
Lucidor (http://lucidor.org/lucidor/) is a good Linux ePub desktop viewer, the best I have tried.

susan_cassidy
02-15-2011, 12:04 PM
Kindle was using a version of Mobipocket format before ePub got so popular. There has been no real reason for them to change.

Ortep
02-16-2011, 01:59 AM
What do you mean by poorly implemented? Most readers do it reasonably well. I agree that there are not a lot of readers there, but then again, the format is more for portable devices like readers.

The implementation is horrific. Standard? What standard? I have three readers and of course the desktop. I have books that look completely different on all of them. Sometimes good but different, sometimes almost unreadable. To display them the same on every reader I'll have to convert them seperately for every reader and edit that arcane CSS file. Most of the time I simply delete it. Books look a lot better that way. There is also absolutely no way to edit the ePub. Sigil is a nice try, but it is nowere near WYSIWYG.


So me and my wife went back to mobi. Not as fancy as ePub, but always the same en what more do you need for a normal book? Italics, font sizes, fat print, pictures and that's it.

Toxaris
02-16-2011, 03:17 AM
Sorry, I don't agree. Why would you need a WYSIWYG? Sigil is not trying to be one. Editing is quite easy with Sigil.

The main problem is usually NOT the readers, but the implementation of the ePub itself. A lot of books use a dirty way of creating ePubs. If you would create an ePub as it should with good clean xhtml and css, it would look almost the same on every reader. Especially if you just use font sizes, italics and alike.
One thing that is dependent on readers is whether justification is done.

Sometimes I take a look at an bought ePub and I cry of what they did to mutilate the book. After editing the book looks as it should, is smaller and faster. Usually editing takes me up to half an hour at most.

MrPLD
02-16-2011, 03:24 AM
WYSIWYG is rather "pointless" overall for ePub, because unless you have an emulator for each and every device you're planning to have viewing the ePub file you'll never be guaranteed of the "perfect rendering".

You have to rather "let go" of the desire to control things at the pixel level else you will go mad. It's something a lot of web-designers really lost a lot of hair over (and still do). I guess that's in part why I like things like LyX (http://www.lyx.org) a lot, because it forces you to let go. So long as you're stylistically consistent should come out quite okay (and if not, then a few CSS tweaks should fix it).

Paul.

pdurrant
02-16-2011, 03:25 AM
So me and my wife went back to mobi. Not as fancy as ePub, but always the same

You would have thought so, wouldn't you? But that's not true. The different physical Kindles render Mobipocket files differently to each other, and Kindle for PC and Kindle for Mac render them differently again.

It's really very disappointing.

MrPLD
02-16-2011, 03:33 AM
It's really very disappointing.

While in one way, yes, it is a bit dissapointing, in another it's an example of the power of the flexibility of the format. Your ePub should be rendered reasonably well irrespective of the display or engine on the device, that is the power of these formats, this means that manufacturers can offer a whole swathe of devices and not be limited to a set of output formats that'll inevitably become outdated within the next generation or two of eReaders.

A very good example of what happens when you try to hard-code the formats is what we have when trying to view PDFs on eReaders. I realise people are using reflowable attributes in PDF but really it's like rounding the edges off a rectangular stone to use as a rolling pin.

Paul.

pdurrant
02-16-2011, 03:58 AM
While in one way, yes, it is a bit dissapointing, in another it's an example of the power of the flexibility of the format.

No, you have not understood me. I did not mean that Mobipocket ebooks rendered differently on different devices because of different font or screen sizes leading to different number of character per page and different line breaks. This is what one expects from ebooks, and is one of their great advantages (which PDFs don't have).

No, I mean that Mobipocket ebooks do not render in the same way on different devices. Coding that works one way on one device works a different way or is ignored on a different device.

Amazon have not made their various different implementation render Mobipocket format ebooks consistently.

MrPLD
02-16-2011, 04:21 AM
My appologies for that - okay, so the mobi format itself is the neutral party, rather Amazon's Kindle rendering engine has the disparity, thanks.

Paul.

DSpider
02-16-2011, 08:41 AM
Isn't XML a more restrictive language ? In other words, isn't it more uptight about closing tags ? Because I thought that with HTML you can leave open tags, which used to cause 99% of the problems with website layouts...

However, today most modern browsers account for that and web page looks more or less the same. And with tools such as epubchecker and Sigil's "Validate Epub" tool you'd think the code would be more precise. Yet they look different everywhere you look at it.

Why can't they focus on developing an open source engine that reads it properly instead of focusing on adding more features (such as audio and video) ?

Ortep
02-16-2011, 09:23 AM
You would have thought so, wouldn't you? But that's not true. The different physical Kindles render Mobipocket files differently to each other, and Kindle for PC and Kindle for Mac render them differently again.

It's really very disappointing.

I don't care about small diffences. But if a mobi book is readable on my Kindle it is also readble on my wife's Bebook and our old Cybook.

That is not the case with ePub on the Bebook and the Cybook and my desktop. The horrible inplementations of ePub sometimes makes borders of 2 cm on one device and no borders at all on the other. Or one device has an indentation of one caracter at every paragraph while the other has something like half a line. Or an extra whilte line in between.

Or on one device everything looks OK while on the other al kinds of (inverted) ??? appear. Convert the file to mobi and most of your troubles are gone. Especially if you rip uit de CSS first.

I agrea that ePub is more sophisticated than Mobi. But everybody implements only 70% of the standard. And by preference an other 70% than the rest of the world.

The end result is that only 50% is common. That is why it looks often better if you rip out the CSS, then none of the fancy stuff that is rendered badly is available anymore.

Sigiel is only good for cutting a file in chapters, but that is something I have no need for unless the text of the file is OK.

Sometimes it works to convert the ePub with Calibre to RTF and edit all the fancy junk out with Word and the reconvert it.

Toxaris
02-16-2011, 02:38 PM
If you say that Sigil is only good for cutting a file in chapters, you do it grave injustice. It has some flaws, sure. But in its right it is a very good tool to create epub's. And yes, without much problems your books look almost the same anywhere.

The examples you mention exactly proves my point. It is not the viewers that are at fault, but the creators of the epub's. They don't follow normal rules and you cannot expect that all viewers to adjust to all steps outside the format. Using pt or px instead of em for example.
Those things are quickly fixed. Much quicker than converting to mobi or RTF. I think it is funy that you remove 'junk' with Word and then reconvert it. There is even more junk in that way...

However, if it works for you.... But put the blame where it should be, the creators of the ePubs.

DSpider
02-16-2011, 03:54 PM
You mean the creators/publishers don't put their ePubs through validators ? Because I think they do, and the files pass. Yet they still display differently on every system. Why is that ? Why can't they develop ONE, GOOD, reading engine and adapt it to multiple devices ? Or even use one already developed:

http://en.wikipedia.org/wiki/Gecko_%28layout_engine%29

Open source, cross platform... What are they, stupid ?

theducks
02-16-2011, 08:40 PM
You mean the creators/publishers don't put their ePubs through validators ? Because I think they do, and the files pass. Yet they still display differently on every system. Why is that ? Why can't they develop ONE, GOOD, reading engine and adapt it to multiple devices ? Or even use one already developed:

http://en.wikipedia.org/wiki/Gecko_%28layout_engine%29

Open source, cross platform... What are they, stupid ?

A Valid Junk file just passed all the Syntax rules :(..
Many have more bloat than a US Gov. budget :eek:
But all the i's have been dotted and t's crossed ;)

The book can look like dog doo and pass the check :rofl:

Toxaris
02-17-2011, 04:48 AM
Indeed, validators only do so much. The engine you mention is not usable. The reason is simple. It is a browser engine, not an ePub engine.
Some specifications in the ePub format are not in a browser engine and vice versa.

They are not stupid, but just don't want to do it. Apparently it is not as simple as it seems.

I would not be surprised if some publishers do not use validators. Don't forget, most publishers are not that happy with e-books. It is an afterthought and the larger publishers would love to see it fail.
There are only two validators at present. The first one is epubcheck. Do a search on the forum. You can really pass garbage through that one without any comment.
The other one is FlightCrew. That one is better, but I can imagine that also through that one garbage can be validated.

If you create a 'normal' book, following the standards and best practices as can be found all around the forum, it will look almost the same on most readers. What do I mean with a 'normal' book? Simple: cover, TOC, headings, paragraphs (with or without indents), dropcaps perhaps, some images, simple formating, things like that. Should serve most books.
Where do problems arise? From what I have seen centering images, floats and more obscure options in CSS. By obscure I mean less used. The centering images is the most stupid one. That one happens on the i-devices and aldiko for sure.

As long as the reader supports CSS, it should look the same.

Jellby
02-17-2011, 07:30 AM
As long as the reader supports CSS, it should look the same.

Except that different readers may have different defaults, and support a different set of the standard CSS (some properties are "optional" in the spec, and it says that reading systems "may" do this or that).

eping
02-17-2011, 07:32 AM
Yet they still display differently on every system. Why is that ? Why can't they develop ONE, GOOD, reading engine and adapt it to multiple devices ? Or even use one already developed:
http://en.wikipedia.org/wiki/Gecko_%28layout_engine%29
Open source, cross platform... What are they, stupid ?

I think the main reason is commercial.

1 Adobe fears ePub will encroach its benefit from PDF.
So they want to keep ePub only in the field of e-ink devices.

2 Microsoft continues their numbness on ebook market.
MS hasn't improved their CHM format for over ten years, actually
CHM is very similar to ePub.

3 ePub committee adopts XHTML instead of HTML.
This make desktop eReaders based on webbrowser meaningless.
For these Readers will show invalid XHTML as HTML correctly,
while such invalid XHTML in ePub will not show properly on eReader
devices.
It's not easy for desktop Readers to simulate the backwardness
of those ebook devices, especially ADE system.
So users will not accept a more robust epub reader on desktop.

4 Before Apple adopted ePub, ePub was unimpressive.
But Apple is not so influential on PC.

So ePub lacks a real leading giant till now.

Toxaris
02-17-2011, 07:56 AM
Apple has not adopted ePub. They are trying to corrupt it. Instead of trying to change the format via the IDPF, the create their own version.
Sure, some changes are good, but there are also bad ones. Apple should call their version iPub or alike and should be forbidden to say they support/adopt ePub.

Toxaris
02-17-2011, 07:58 AM
Except that different readers may have different defaults, and support a different set of the standard CSS (some properties are "optional" in the spec, and it says that reading systems "may" do this or that).

Agree, the specs are not perfect. However, the most used properties are implemented ok. The defaults, that is another question all together...