Minimal and ideal EPUB3 testcase?

[Doitsu]

Originally Posted by ibu

I looked at the nav.xhtml of the book "The forerunner" of the link you posted in the last #12. It's a bit far from minimal.

It might look a bit convoluted, but it contains just enough epub3 markup to pass epubcheck.

The easiest way to find out, what tags are required is to delete tags and run epubcheck afterwards. (If you use the Sigil epubcheck plugin, you don't have to save files after deleting/adding tags, because the plugin will automatically grab the latest files.)

Originally Posted by ibu

1 When the charset is declared in the xml element, is
necessary to declare it a second time in the meta element

AFAIK, they're optional.

Originally Posted by ibu

2 Do we need to language declarations?

AFAIK, they're optional.

Originally Posted by ibu

3 Do we need the CSS styles? Are the ebook devices not able to show a nice TOC without it?

CSS styles are not required and most epub3 apps will ignore them. (These styles are intended for epub2 apps.)

Originally Posted by ibu

4 Should we place the heading "Table of content" inside the title element or in a h1?

Since <title></title> tags are ignored by most epub apps, you might want to put Table of Contents in a heading tag. BTW, heading tags (h1-h6) are optional in NAV docs.

Originally Posted by ibu

5 For what are the attributes in the body and nav element useful?

Most of these attributes are mandatory in epub3 content docs. For details, see the epub3 specs.

[ibu]

@BetterRed
Thx for the hint. I changed to "code", than "noparse" is not necessary anymore.

[ibu]

I started as you, Doitsu, suggested - with installing the epubcheck plugin for Sigil.

I'm getting nearer, slowly.

My actual OPF:

<?xml version="1.0" encoding="utf-8"?>
<package 
version="3.0" 
unique-identifier="book-id" 
xmlns:dc="http://purl.org/dc/elements/1.1/" 
xmlns="http://www.idpf.org/2007/opf">
  <metadata>
    <dc:title>Tanz der seligen Geister</dc:title>
    <dc:creator>Alice Munro</dc:creator>
    <dc:language>de-DE</dc:language>
    <dc:publisher>Dörlemann</dc:publisher>
    <dc:date>2012</dc:date>
    <dc:identifier id="book-id">urn:isbn:978-3-908778-02-8</dc:identifier>
  </metadata>
  <manifest>
    <item id="title" href="Text/title.xhtml" media-type="application/xhtml+xml"/>
    <item id="content" href="Text/content.xhtml" media-type="application/xhtml+xml"/>
    <item id="about" href="Text/about.xhtml" media-type="application/xhtml+xml"/>
    <item id="imprint" href="Text/imprint.xhtml" media-type="application/xhtml+xml"/>
    <item id="nav" href="Text/nav.xhtml" media-type="application/xhtml+xml" properties="nav"/>
    <item id="default.css" href="Styles/default.css" media-type="text/css"/>
    <item id="author.jpg" href="Images/author.jpg" media-type="image/jpeg"/>
    <item id="cover.jpg" href="Images/cover.jpg" media-type="image/jpeg"/>
  </manifest>
  <spine>
    <itemref idref="title"/>
    <itemref idref="content"/>
    <itemref idref="about"/>
    <itemref idref="imprint"/>
    <itemref idref="nav"/>
  </spine>
</package>

Open questions:

1 When I've started the plugin, the function "Edit > Undo" is not available.
That makes it difficult to test all the possible mimilizations. Is there are good way to deal with it?

2 How can I delete the file toc.ncx in Sigil permanently please? The context menu does not offer a function to delete it.
When I delete <item id="ncx" href="toc.ncx" media-type="application/x-dtbncx+xml"/> in the OPF, it is automatically inserted again by Sigil, when I reopen the file. How can I prevent that?

3 How can I remove
<meta name="Sigil version" content="0.9.7" />
<meta property="dcterms:modified">2017-03-07T12:14:04Z</meta>
permanently?

4 Can I place the files in the main level or is it impossible in Sigil to delete the directories "Text", "Images" and "Styles"?

5 The code above validates almost. The output of the plugin:

File: content.opf
Line: 7
Offset: 188
Message: Col: 13: ERROR(RSC-005): Error while parsing file 'package dcterms:modified meta element must occur exactly once'.

The next step in the minimization will be to check, if that OPF works with a Kobo.
I'm frightened

[Doitsu]

Originally Posted by ibu

1 When I've started the plugin, the function "Edit > Undo" is not available.

Plugin changes aren't saved in the Undo buffer. What I meant is that you could load a book delete/add a tag and run epubcheck afterwards without having to save the book in between changes.

For example, you could create a new epub3 book, add a single <p> tag after the first paragraph

<p>&#160;</p><p>

and run epubcheck, which will display the following error:
'

FATAL(RSC-016): Fatal Error while parsing file 'The element type "p" must be terminated by the matching end-tag "</p>".'.

If you then remove the single <p> tag and re-run epubcheck the error message will go away, even though you did't save the book.

Originally Posted by ibu

2 How can I delete the file toc.ncx in Sigil permanently please?

You can't delete toc.ncx via the GUI, because Sigil tries to generate epub2 compatible epub3 books; you'll have to unzip the epub and remove toc.ncx and the corresponding .opf file entry and zip it up again.
(You might want to use the ePub Zip/Unzip AppleScript application for Mac OS X.)

Originally Posted by ibu

3 How can I remove
<meta name="Sigil version" content="0.9.7" />

You'll have to unzip the epub and manually delete it.

Originally Posted by ibu

3 How can I remove
<meta property="dcterms:modified">2017-03-07T12:14:04Z</meta>
permanently?

This entry is required for all valid epub3 books and shouldn't be removed.

Originally Posted by ibu

5 The code above validates almost.

See previous answer.

Originally Posted by ibu

4 Can I place the files in the main level or is it impossible in Sigil to delete the directories "Text", "Images" and "Styles"?

Sigil will automatically move all files to folders based on their mime types. AFAIK, Calibre Editor will maintain the original folder structure, but it only has minimal epub3 support.

[ibu]

I continued with the minimization of the nav.xhtml.

The following code validates with the plugin:

<?xml version="1.0" encoding="utf-8"?>
<!DOCTYPE html>
<html xmlns="http://www.w3.org/1999/xhtml" xmlns:epub="http://www.idpf.org/2007/ops" xml:lang="de">
<head></head>
<body>
<h2>Inhaltsverzeichnis</h2>
<nav epub:type="toc">
	<ol>
		<li><a href="../Text/main.xhtml#t1">Der Walker Brothers-Cowboy</a></li>
		<li><a href="../Text/main.xhtml#t2">Die leuchtenden Häuser</a></li>
		<li><a href="../Text/main.xhtml#t3">Bilder</a></li>
		<li><a href="../Text/main.xhtml#t4">Danke für die Schlittenfahrt</a></li>
		<li><a href="../Text/main.xhtml#t5">Das Büro</a></li>
		<li><a href="../Text/main.xhtml#t6">Ein Gläschen Medizin</a></li>
		<li><a href="../Text/about.xhtml#t7">Zur Autorin und zu ihrer Übersetzerin</a></li>
		<li><a href="../Text/about.xhtml#t8">Zum Buch</a></li>
		<li><a href="../Text/imprint.xhtml#t9">Impressum</a></li>
	</ol>
</nav>
</body>
</html>

Open questions:

1 Is epub:type="toc" in the nav element mandatory?
Until now I couldn't find it in the specs you mentioned: http://www.idpf.org/epub/31/spec/epub-contentdocs.html
Independent of that, Sigil deletes automatically the complete content of the nav element, when one deletes epub:type="toc"

2 How can I teach Sigil to produces such a minimized nav.xhtml

3 Will that navigation work with a Kobo?

[Doitsu]

Originally Posted by ibu

1 Is epub:type="toc" in the nav element mandatory?

AFAIK, no. You'll have to ask KevinH about this. IIRC, it was added to indicate that the machine readable NAV also doubles as a human readable TOC, but I might be wrong. Either way removing this attribute won't achieve anything.

Originally Posted by ibu

2 How can I teach Sigil to produce such a minimized nav.xhtml

AFAIK, there are no options for customizing the NAV doc. You'll have to unzip the epub and manually remove elements that you don't like. However, it'd be rather pointless, because removing them won't make your book load faster.

Originally Posted by ibu

3 Will that navigation work with a Kobo?

I've got no idea. Theoretically, I should work if you also change the file extension to .kepub.epub. You might want to ask about this in the MR Kobo forum.

[ibu]

@Doitsu as reply to your post #19

Thanks again for your time and all the helpful hints.

I understood how to deal with "undo" now.

Great tip with the zip/unzip App. Works perfectly.
I deleted the NCX file and deleted the appropriate references in the OPF.
Than I zipped everything with the app: it validates with http://www.pagina-online.de/software/epub-checker/
But when I open the EPUB in Sigil, it automatically adds a NCX.
Is there any way to prevent that in Sigil?

Ditto for the entry about the Sigil version in the meta element.

About the folder structure:
OK, I like Sigil a lot and want to use it for editing. So I have to live with the fixed folder structure.

I reached the target of a minmal epub almost - mainly because of your hints

Well, I need to test, what the Kobo will do with such a minimal epub3.

I hope there will be a solution for Sigil to keep it prevent that automatically generated NCX.

Are there strong reasons against the idea to place the complete content into one single file (title, content, about, imprint)? Will that cause problems in a Kobo? I think of inserting automatical page breaks in front of specific headings via CSS. Than there's no need at all for separated files for the content.

[ibu]

@Doitsu as reply to #21

I tried to validate with and withou epub:type="toc".
Result: It is mandatory.

About the minimized nav:
What a pity, that there's no way to create it with Sigil.

[pdurrant]

Originally Posted by BetterRed

fyi - you can eliminate the :p smileys by wrapping the text in [noparse][/noparse] tags.

BR

Or check check the "Disable smilies in text" checkbox in the Advanced editing/composing options.

:) Like I've done for this post.

[Doitsu]

Originally Posted by ibu

Then I zipped everything with the app: it validates with http://www.pagina-online.de/software/epub-checker/

BTW, this software use the same validator as the epubcheck plugin.

Originally Posted by ibu

But when I open the EPUB in Sigil, it automatically adds a NCX. Is there any way to prevent that in Sigil?

No. You'll have to use Calibre Editor for final edits. Make sure to install the Calibre epubcheck plugin and run it before and after your edits. However, Calibre might add its own set of custom metadata entries, unless you disable the following option:

Edit > Preferences > Integration with Calibre > Update metadata ...

Originally Posted by ibu

Ditto for the entry about the Sigil version in the meta element.

You can't suppress this metadata entry, which is useful when tracking down version specific bugs.

Originally Posted by ibu

I hope there will be a solution for Sigil to keep it prevent that automatically generated NCX.

KevinH has indicated that future versions might have an option to suppress the automatic generation of NCX files, however, you might want to keep the NCX file for backwards compatibility with epub2 apps.

Originally Posted by ibu

Are there strong reasons against the idea to place the complete content into one single file (title, content, about, imprint)? Will that cause problems in a Kobo? I think of inserting automatical page breaks in front of specific headings via CSS. Than there's no need at all for separated files for the content.

If you use separate files, they'll usually load faster. Also some older epub apps have problems with files larger than 300-400KB.
As for Kobo-specific questions, you'll have to ask them in the Kobo forum.

[ibu]

I minimalized the file structure and I even eliminated the nav.xhtml and integrated the code into content.xhtml.
At the moment I think, there's no way to enhance the minimalization.
There's no other epub3 in the world, which could beat it


The attached epub3 validates. Please do not open it in Sigil, if you like to look into it, because Sigil will destroy the file structure.

Tomorrow I have to check out, if such a minimal epub3 works with a Kobo.
I will give a feedback here.

content.opf

<?xml 
version= "1.0" 
encoding="utf-8"?>
<package 
  version=          "3.0" 
  unique-identifier="book-id" 
  xmlns:dc=         "http://purl.org/dc/elements/1.1/" 
  xmlns=            "http://www.idpf.org/2007/opf">
  <metadata>
    <dc:identifier id="book-id">      urn:isbn:978-3-908778-02-8</dc:identifier>
    <meta property="dcterms:modified">2017-03-07T12:14:04Z      </meta>
    <dc:language>                     de-DE                     </dc:language>
    <dc:title>                        Tanz der seligen Geister  </dc:title>
    <dc:creator>                      Alice Munro               </dc:creator>   <!-- optional -->
    <dc:publisher>                    Dörlemann                 </dc:publisher> <!-- optional -->
    <dc:date>                         2012                      </dc:date>      <!-- optional -->
  </metadata>
  <manifest>
    <item id="content" href="content.xhtml" media-type="application/xhtml+xml" properties="nav"/>
    <item id="cover"   href="cover.jpg"     media-type="image/jpeg"/>
  </manifest>
  <spine>
    <itemref idref="content"/>
  </spine>
</package>

content.xhtml

...
<h2>Inhaltsverzeichnis</h2>
<nav epub:type="toc">
	<ol>
		<li><a href="content.xhtml#t1">Der Walker Brothers-Cowboy</a></li>
		<li><a href="content.xhtml#t2">Die leuchtenden Häuser</a></li>
		<li><a href="content.xhtml#t3">Bilder</a></li>
		<li><a href="content.xhtml#t4">Danke für die Schlittenfahrt</a></li>
		<li><a href="content.xhtml#t5">Das Büro</a></li>
		<li><a href="content.xhtml#t6">Ein Gläschen Medizin</a></li>
		<li><a href="content.xhtml#t7">Zur Autorin und zu ihrer Übersetzerin</a></li>
		<li><a href="content.xhtml#t8">Zum Buch</a></li>
		<li><a href="content.xhtml#t9">Impressum</a></li>
	</ol>
</nav>
</body>
</html>

[DaleDe]

I can't see the cover image. Otherwise it looks good. There is a toc at the end of the document. It is visible in ADE. I would be better near the front.

Dale

[Doitsu]

Originally Posted by DaleDe

I can't see the cover image.

The cover image entry in the .opf file is missing a properties="cover-image" attribute:

<item id="cover.jpg" href="cover.jpg" media-type="image/jpeg" properties="cover-image"/>

@ibu: Even though epubcheck didn't flag this as an error, the Nav doc needs to be a separate .xhtml file. What you did is create a Nav doc with book contents, which is not what you want.

You also might want to download Azardi and ADE 4.5 and test your book with it.

[ibu]

Thanks for the hint about the cover image.
I added the attribute "properties".

And I separated the the nav.

The validated attachement contains the changes. When we finishe this thread, I will edit the first posting and will attache the final file so it is easier for readers to find the latest version.

The cover is not visible in the Kobo. Nor it is the TOC.
But I haven't started with the work to make the minimal epub3 compatible to KOBO. That may be the reason.

Is it true, that the nav must appear inside the spine element? It couldn't find information about that in the specs.

About the validator:
May be it would be useful, when it warns, if a nav does not exist as a separated file?

[Doitsu]

Originally Posted by ibu

The cover is not visible in the Kobo.

The cover is shown in Azardi and iBooks.

Originally Posted by ibu

Nor it is the TOC.

Most likely the book wasn't opened with the ACCESS NetFront™ engine and RMSDK doesn't support NAV docs.

As I've repeatedly pointed out in this thread, you'll need to change the file extension to .kepub.epub and you also might need to implement other KEPUB specific changes.