Feedbooks to support ePub format

[Bob Russell]

Fans of Feedbooks will be happy to learn that the new ePub standard will soon be supported in addition to PDF A4, Sony Reader, iLiad and custom PDF.

Teleblog is reporting the following tidbit... "'Feedbooks will have on-the-fly .epub files this month,' reports Hadrien. 'We’ll enable this feature on books first and then on RSS feeds too. In the future, we’ll work on improving the overall look of these .epub files and would also like to add a ‘custom .epub’ feature on the website, where anyone will be able to easily customize the CSS and layout of the book.'

From the Feedbooks site, "Feedbooks is a project in development in Paris, France. Our main focus is to create a complete experience for e-ink readers and other portable reading devices... Made with e-paper devices in mind (Sony PRS-500 & iRex iLiad), Feedbooks is a complete experience for mobile reading. Anyone can easily publish content, and customize the formatting."

[Hadrien]

Still a work in progress, but here's a sample file that our development website can currently generate. Might have a few bugs, metadatas and the footnotes are missing, but overall it works quite well already and you get a pretty good idea of what our .epub files will look like.

Expect the epub output to be available as a public beta this week.

PS: Could Alexander or any admin add epub support for the uploads ?

[ricdiogo]

Can't open it using Digital Editions Hadrien. I see the TOC but when I click the chapters I can't read them.

[Hadrien]

Originally Posted by ricdiogo

Can't open it using Digital Editions Hadrien. I see the TOC but when I click the chapters I can't read them.

Oh I've tried it and it works, but Adobe DE is a very slow and inefficient software, and War & Peace is 350+ chapters. I'll replace it with something lighter.

[wallcraft]

Having .epub extension files available for download would help with web-based epub readers like OpenBerg Lector (and Digital Editions for that matter).

There is a real need for a ePub compliance checker. I'm not sure Dostoyevsky-The_Gambler is a valid .epub. It fails using Lector for example. FBReader can read it, but it currently lacks TOC support which is a bit limiting.

[Hadrien]

Originally Posted by wallcraft

Having .epub extension files available for download would help with web-based epub readers like OpenBerg Lector (and Digital Editions for that matter).

There is a real need for a ePub compliance checker. I'm not sure Dostoyevsky-The_Gambler is a valid .epub. It fails using Lector for example. FBReader can read it, but it currently lacks TOC support which is a bit limiting.

Well, we don't have any validation tool yet, so we created our files based on the examples provided by IDPF and the docs mostly.
They work in DE, although DE is kinda slow if your .xml file is too big. Haven't tried yet with FBReader but FBReader currently lacks both TOC and CSS support which is quite limiting.
We're correcting a few things and we'll add this output on the website for beta testing (I guess today or tomorrow).

Keep in mind that this is pure beta testing for the moment, if you find any problem with the files, we'll correct them.

PS: Just tried OpenBerg and found a small bug in our file. Corrected the whole thing manually and now it opens perfectly in OpenBerg.

[Hadrien]

Ok. Here's another screenshot, in OpenBerg this time.

I can already notice a problem with DE: I had to put my paragraphs in <p></p> but also add a <br /> tag between them for DE. In OpenBerg, it is displayed correctly, which mean that I get blank lines (that I don't want in my file).

[DaleDe]

Originally Posted by Hadrien

Well, we don't have any validation tool yet, so we created our files based on the examples provided by IDPF and the docs mostly.
They work in DE, although DE is kinda slow if your .xml file is too big. Haven't tried yet with FBReader but FBReader currently lacks both TOC and CSS support which is quite limiting.
We're correcting a few things and we'll add this output on the website for beta testing (I guess today or tomorrow).

Keep in mind that this is pure beta testing for the moment, if you find any problem with the files, we'll correct them.

PS: Just tried OpenBerg and found a small bug in our file. Corrected the whole thing manually and now it opens perfectly in OpenBerg.

I just tried The Gambler (I don't know where OpenBerg is). I looked at it in digital editions and I converted it to my eb1150 as a test. Do you mind if I finish the conversion to eb1150 and then post it at mobileread? Here are some comments.

1. The digital editions is really slow to bring up the document.
2. I was surprise the zip container had no compression. Was this done to try and increase speed? Normally you should be able to publish with the .epub extension with full zip compression.
3. The file extensions were xml instead of xhtml. Is there a reason for this? It might slow down the processing but I am not sure.
4. The note.xml file was included but I do not think it was referenced anywhere.
5. The final feedbooks page has a very light graphic. Did this come across ok in the readers you are targeting? It looked too light when converted to 16 level gray scale on my device. I had to darken it a bit.
6. The TOC did not come across to my eb1150 version but I can regenerate it manually. It is probably a limit in my system. It worked ok in digital editions.
7. There are a lot of " marks with space on both sides in the document. This is one of the nits that I hate. I corrected it in my copy of the xml file and can send it to you if you wish.
8. Are you planning to fix the quotes to curly quotes?
9. There are some characters in the file that my reader did not recognize but that may just be my reader.

I hope this helps.
Dale

[Hadrien]

Originally Posted by DaleDe

I just tried The Gambler (I don't know where OpenBerg is). I looked at it in digital editions and I converted it to my eb1150 as a test. Do you mind if I finish the conversion to eb1150 and then post it at mobileread? Here are some comments.

1. The digital editions is really slow to bring up the document.
2. I was surprise the zip container had no compression. Was this done to try and increase speed? Normally you should be able to publish with the .epub extension with full zip compression.
3. The file extensions were xml instead of xhtml. Is there a reason for this? It might slow down the processing but I am not sure.
4. The note.xml file was included but I do not think it was referenced anywhere.
5. The final feedbooks page has a very light graphic. Did this come across ok in the readers you are targeting? It looked too light when converted to 16 level gray scale on my device. I had to darken it a bit.
6. The TOC did not come across to my eb1150 version but I can regenerate it manually. It is probably a limit in my system. It worked ok in digital editions.
7. There are a lot of " marks with space on both sides in the document. This is one of the nits that I hate. I corrected it in my copy of the xml file and can send it to you if you wish.
8. Are you planning to fix the quotes to curly quotes?
9. There are some characters in the file that my reader did not recognize but that may just be my reader.

I hope this helps.
Dale

Hello Dale,

1) I noticed that we had a few tags that were not correctly closed, that's why it wasn't working on OpenBerg but only 1-2 were concerned and not in the main text. DE parser seems to be really bad, no speed on both OpenBerg and FBReader.
2) You're not supposed to compress the files, zip is just a container here. Some software might support compression, but it might slow down the whole thing and we have to make those files fully compliant (that's what a beta test is for...).
3) I don't think that .xml or .xhtml changes anything.
4) We left a few files from the example we used to create our files (myantonia), that's why there's a few metadata and files left. We're removing all references now and cleaning the whole thing (I took this file while developing the output).
5) The graphic looked ok on my Sony Reader. Might darken it a bit though...
6) TOC worked perfectly with all the files I've tested but only DE seems to support TOC.
7) This is how the text itself was uploaded, I guess we could add another layer of formatting though, and this is a constant work in progress.
8) Same here, if we could easily identify the curly quotes we could add this. In some situation, it's hard to automatically fix this.
9) Haven't seen any unusual characters using DE or OpenBerg. Might be linked to the fact that ePub is fully UTF-8 ?

Post any conversion that you want to. Did you manually converted the file ? Automatically ? Which output are you using ?

Thanks for the feedback,
I'll post an updated file soon.

[DaleDe]

Originally Posted by Hadrien

Hello Dale,
Post any conversion that you want to. Did you manually converted the file ? Automatically ? Which output are you using ?

Thanks for the feedback,
I'll post an updated file soon.

I manually converted the file but it wasn't much effort. I have been exploring the need to support epub and in spite of what the technical releases and publications put out on epub ebook technologies does not directly support epub on any existing products. They probably will on some new version that they have in their labs but not on anything they have released.

Here is what I did to convert the file.
1. unzipped the container.
2. created my own opf file as ebook publisher 2.2.5 won't read an epub one.
3. renamed all the extensions from .xml to .odf (html would also have worked).
4. imported the documents and arranged them. It then compiled ok with some warnings.
5. corrected css files that import other files. My tools will not accept quotes around the imported filenames.

It compiled ok with a few known warnings on unsupported structures. ebook publisher is back in the dark ages for support of css.

I will need to build a toc to make it releasable by me but it looks good so far. I also need to correct some spacing problems on the title page but the data is ok.

My chapter titles do not have the box around them but otherwise my book looks similar to the one you produced. My paragraphs are not indented but I can fix this.

Dale

[wallcraft]

Originally Posted by Hadrien

You're not supposed to compress the files, zip is just a container here.

Everything EXCEPT the first file (mimetype) can be, and will typically be, compressed. Any reader that supports ePub will support compressed content. The requirement that mimetype not be compressed is so that programs like Linux's file command can recognize the type by looking at a file's first few bytes. See the .epub format thread for more useful links.

[Hadrien]

Originally Posted by wallcraft

Everything EXCEPT the first file (mimetype) can be, and will typically be, compressed. Any reader that supports ePub will support compressed content. The requirement that mimetype not be compressed is so that programs like Linux's file command can recognize the type by looking at a file's first few bytes. See the .epub format thread for more useful links.

OK, I'll switch this to compression except for the mimetype then. Files will be much smaller this way !

[JSWolf]

Originally Posted by Hadrien

Ok. Here's another screenshot, in OpenBerg this time.

I can already notice a problem with DE: I had to put my paragraphs in <p></p> but also add a <br /> tag between them for DE. In OpenBerg, it is displayed correctly, which mean that I get blank lines (that I don't want in my file).

Does the .epub books need those line spaces between paragraphs? If so, I really won't be reading them as I find them rather annoying on a large enough screen. Can the quotes and apostrophes be converted to the curly kind?

[Hadrien]

Originally Posted by JSWolf

Does the .epub books need those line spaces between paragraphs? If so, I really won't be reading them as I find them rather annoying on a large enough screen. Can the quotes and apostrophes be converted to the curly kind?

I don't like blank lines either. Take a look at the DE screenshots that I posted.

I had to do this to get this result on DE: <p>text</p><br /><p>text2</p>

Of course, because of this, I get blank lines in OpenBerg...

Curly kind once again can be done two different ways: people uploading the text already include curly quote, or we could add some extra pre-processing on the text.

[andym]

It all seems to be working fine on DE for me (I'm on OS X though that shouldn't make a difference).

It's late, and no doubt I'm missing something but I don't understand the problem with the spaces between paragraphs problem. Most of the epubs I have use the normal html syntax (ie no break tags), and dsiplay as they should - ie the paragraphs flowing one after another without white space. The one exception is White Fang which uses CSS styles to give an extra margin at the beginning of paragraphs. If you need to have spaces between paragraphs that seems the better way to do it.

If you have any problems or issues with DE it would be well worth posting to the Adobe DE forum - in my experience the project engineers do seem to read and respond to posts and I'm sure they'd be happy to help.

Dale

OpenBerg is here (on Sourceforge)

http://openberg.sourceforge.net/