Endnotes within Endnotes

[Alcuin7]

I’m a newbie. I’m probably asking a question that’s been asked before, but I was unable to locate it using the forum Search or Google.

I have document with endnotes embedded to endnotes. (Footnotes to footnotes.) That is, some of the endnotes have their own endnote references. In at least some instances, there are endnotes to endnotes to endnotes.

In Microsoft Word, I overcame this using cross-links. All the embedded (secondary and tertiary) endnotes are listed in main document following the primary endnote, then shrunk to font size 1 and colored white to make them as inconspicuous as possible. In the primary endnote, I cross-linked to each endnote.

I converted my DOCX to EPUB with Calibre, but it doesn’t recognize the cross-links. And the “disappeared” endnotes show up in the main text.

So I located a secondary endnote in the converted EPUB and used the Calibre HTML editor to cut-and-paste that secondary endnote from the main text into the primary endnote where I want it. That worked great: I can go to the primary endnote, click on the secondary endnote, and I go to the secondary endnote. So far, so good.

Now my problem arises. I click on the secondary endnote number (in the secondary endnote) to return to the primary endnote, and instead of going back to the primary endnote, I go back to the main text.

Does anyone have any idea about what I need to do to return to the primary endnote instead of the main text?

[Tex2002ans]

Originally Posted by Alcuin7

I have document with endnotes embedded to endnotes. (Footnotes to footnotes.) That is, some of the endnotes have their own endnote references. In at least some instances, there are endnotes to endnotes to endnotes.

This type of formatting is very complex.

The only way you will be able to correct this is to get your hands dirty in the actual HTML.

Back in 2013, I posted an EPUB I converted with 3 sets of footnotes:

• *, †, ‡
• (a), (b), (c)
• (1), (2), (3)

Feel free to take that apart and see how I did it.

It had many footnotes in footnotes. For example, Chapter 9:

Within Normal Text:

Spoiler:

<p>External commerce consists of the supply of the home market with foreign, and of foreign markets with home products.<a href="#fn1" id="ft1">[1]</a></p>

Footnotes At End of HTML File:

Spoiler:

<p><a href="#ft1" id="fn1">[1]</a> Products that are bought to be re-sold, are called <i>merchandise</i>; and merchandise bought for consumption is denominated <i>commodities</i>.<a href="#fna" id="fta">[a]</a></p>

<p><a href="#fta" id="fna">[a]</a> This distinction has been discarded in the translation, for the sake of simplification; the general term products being sufficiently intelligible and specific. <span class="smallcaps">Translator</span>.</p>

So the [1] (Red) footnotes point back/forth to each other, and the [a] (Blue) footnotes point back/forth to each other.

You'll have to go through your code, copy/paste them to their proper locations, and come up with all the proper HTML code to link them back and forth.

Originally Posted by Alcuin7

All the embedded (secondary and tertiary) endnotes are listed in main document following the primary endnote, then shrunk to font size 1 and colored white to make them as inconspicuous as possible. In the primary endnote, I cross-linked to each endnote.

... and what happens in the Word DOCX if someone clicked the cross-link? How are people supposed to read these footnotes in footnotes if they're size 1 and white text?

[Alcuin7]

Originally Posted by Tex2002ans

This type of formatting is very complex.

The only way you will be able to correct this is to get your hands dirty in the actual HTML.You'll have to go through your code, copy/paste them to their proper locations, and come up with all the proper HTML code to link them back and forth.

I was already digging into the HTML: that’s why I am here, looking for precisely what you are telling me. And a hearty and sincere “Thank you!” for your guidance!

For instance in the file index_split_006.html there is a primary (first-level) footnote (endnote) 31 with two explanatory secondary (second-level) footnotes (endnotes) numbered 32 and 33. The existing code in index_split_006.html reads (with line-breaks added to make it more legible),

<sup class="block_25">
<sup id="back_note_31" class="calibre10">
<a href="index_split_051.html#note_31" title="31" class="calibre2">
31
</a>
</sup>
</sup>
<sup id="id_Ref489614248" class="text_6">
<sup id="back_note_33" class="noteref1">
<a href="index_split_053.html#note_33" title="33" class="calibre2">
33
</a>
</sup>
</sup>

I moved endnote 32 before my initial post. It is in index_split_006.html as well, but has been cut-and-pasted to replace what had been a cross-link in Microsoft Word embedded into endnote 31 much further down in the html. (Again, line-breaks added)

<sup id="id_Ref489614211" class="text_6">
<sup id="back_note_32" class="noteref1">
<a href="index_split_052.html#note_32" title="32" class="calibre2">
32
</a>
</sup>
</sup>

If I understand what you have written, what I have observed in the code, and the behavior of the EPUB document, when asked to return to the calling point for note 32, the code searches (sequentially?) for “index_split_052.html”, and not finding it in its expected place following index_split_051.html, returns the reader either to that point (the marker for note 31) or to index_split_053.html (the marker for note 33).

1. Are you saying that I need to use another format of endnote markers, say {a, b, c, …}, to identify my second-order notes, and maintain them separately from the first layer of endnotes? I don’t suppose I can begin renumbered them from {a} every time I begin a new endnote with endnotes attached to it? I have 52 sub-notes, and while some of them are tertiary notes (third level), I should run into double-letter note markers (i.e., {aa, ab, ac, …}). This isn’t a problem, but if the tertiary (third level) are numerous – I haven’t counted how many there are – I could be running into four or five asterisks and daggers unless the symbols have their own specified order in Calibre.
2. I assume I must also renumber “index_split_nnn.html” sequentially throughout the document so that they are in their correct order. I think that is clear from your post. There are 454 htmls numbered index_split_000.html to index_split_453.html. I can either export them one-by-one or copy-and-paste them, but that is a real pain for so many pieces of the document. Is there some way to access them all? because if there is, then this becomes a fairly standard search-and-replace hack with external code.
3. I am delighted that you have answered this query, Tex2002ans. Since I know you are also familiar with MOBI, can I straightforwardly convert the EPUB (once I have it) into MOBI without further manipulation? And how do I prevent MOBI from justifying the text, but keep it aligned left?

Many, many thanks! I will begun perusing the HTML in your example EPUB from Jean-Baptiste Say for exemples. You built this with Sigil rather than Calibre? Making the footnote-in-footnote changes manually? Do I need to switch to Sigil? (I don’t want to do that if I can help it.)

Originally Posted by Tex2002ans

...what happens in the Word DOCX if someone clicked the cross-link? How are people supposed to read these footnotes in footnotes if they're size 1 and white text?

You misunderstand. There is no way to embed footnotes into footnotes in Microsoft Word. It is accomplished by entering separate footnotes, formatting and cross-linking them so that they appear to be footnotes to footnotes, then hiding the real footnote markers.

In my Word Document, the embedded footnotes follow immediately after the main footnote, [31][32][33] in this example. In footnote 31, text “32” and “33” are put in their proper positions in footnote 31, superscripted, then cross-linked to the real footnotes 32 and 33. That highlights them and makes them accessible by clicking on them. Then the text for the real footnote markers in the main body of the book (not the footnotes themselves) is reduced to font size 1 and colored white: the real markers are no longer visible. It appears to anyone reading the text that there is nothing there but a clickable marker for footnote 31 (followed perhaps by an unusually wide space), while the secondary “footnotes” appear to be in their proper (cross-linked) places in footnote 31. It looks fine in Word and prints correctly, except for the space behind the footnote marker for [31], which is small enough to escape notice.

But this method does not transpose to EPUB or MOBI after conversion.

[thiago.eec]

Hello, @Alcuin7.

You misunderstood the explanation of @Tex2002ans. It's all about getting the right anchor (the "id" atribute). All you have to do is reference the FILE (.html) and the ID in the link, so it will take you anywhere you like inside your book.

Working with your examples, we have this:

MAIN TEXT FILE - FIRST-LEVEL FOOTNOTE
File: index_split_006.html

<sup class="block_25">
  <sup id="back_note_31" class="calibre10">
    <a href="index_split_051.html#note_31" title="31" class="calibre2">31</a>
  </sup>
</sup>

FIRST-LEVEL FOOTNOTE FILE - SECOND-LEVEL FOOTNOTE
File: index_split_051.html

<sup id="id_Ref489614211" class="text_6">
  <sup id="back_note_31_a" class="noteref1">
    <a href="index_split_052.html#note_31_a" title="31-a" class="calibre2">31-a</a>
  </sup>
</sup>

FIRST-LEVEL FOOTNOTE FILE - SECOND-LEVEL FOOTNOTE
File: index_split_051.html

<sup id="id_Ref489614248" class="text_6">
  <sup id="back_note_31_b" class="noteref1">
    <a href="index_split_053.html#note_31_b" title="31-b" class="calibre2">31-b</a>
  </sup>
</sup>

ATTENTION: Notice that I moved your second-level footnotes references out of the main text file to the first-level footnote file, since those notes do not relate to the main text, but only to that particular note.

Also, notice that the TEXT of you footnotes are located on other files. The footnote 31 text is located on the file index_split_051.html; 31-a is on index_split_052.html; and 31-b is on index_split_053.html.

Now, you must reference each footnote based on their location (the html file) and ID. With that in mind, you should arrange it like this:

TEXT FILE FOR THE FIRST-LEVEL FOOTNOTE
File: index_split_051.html

<p id="note_31"><a href="index_split_006.html#back_note_31">31.</a> TEXT OF THE FOOTNOTE 31. HERE IS THE REFERENCE FOR THE FIRST NOTE: 31-a. LATER ON, THIS IS THE REFERENCE FOR THE SECOND NOTE: 31-b.

TEXT FILE FOR THE SECOND-LEVEL FOOTNOTE
File: index_split_052.html

<p id="note_31_a"><a href="index_split_051.html#back_note_31_a">31-a. </a> TEXT OF THE FOOTNOTE 31-a</P>

Notice that now you will not reference the main text file (index_split_006.html). Instead, you will reference the first-level footnote file (index_split_051.html). Also, the ID is the one for the 31 footnote, since 31-a is related only to that note, and not to the main text.

The same goes for 31-b:

TEXT FILE FOR THE SECOND-LEVEL FOOTNOTE
File: index_split_053.html

<p id="note_31_b"><a href="index_split_051.html#back_note_31_b">31-b.</a> TEXT OF THE FOOTNOTE 31-b</P>

To see this in action, look at the file attached.


EDIT:
Obs.: When you put the ID inside the link element (<a>), Calibre will threat it like a pop-up footnote, wich is great in most cases, but can be problematic in your case, with nested footnotes. To avoid this, I placed the ID inside the paragraph element (<p>) or SUP.

[Tex2002ans]

Originally Posted by Alcuin7

For instance in the file index_split_006.html there is a primary (first-level) footnote (endnote) 31 with two explanatory secondary (second-level) footnotes (endnotes) numbered 32 and 33. The existing code in index_split_006.html reads (with line-breaks added to make it more legible),

If you could attach a sample DOCX/EPUB, that would be helpful.

Originally Posted by Alcuin7

If I understand what you have written, what I have observed in the code, and the behavior of the EPUB document, when asked to return to the calling point for note 32, the code searches (sequentially?) for “index_split_052.html”, and not finding it in its expected place following index_split_051.html, returns the reader either to that point (the marker for note 31) or to index_split_053.html (the marker for note 33).

Let me color code. That may make it easier to understand.

If you understand the basics of HTML, you should be able to parse each chunk.

Clean/Minimalist Footnote

In the text:

<p>External commerce consists of the supply of the home market with foreign, and of foreign markets with home products.<a href="#fn1" id="ft1">[1]</a></p>

Footnote (within the same file):

<p class="footnote">This is an example footnote.<a href="#ft1" id="fn1">[1]</a></p>

The code for a link back/forth is equal and opposite.

Text: id = ft1, points to fn1.
Footnote: id = fn1, points to ft1.

Endnote

Since you have to point to a completely different file, the code gets a little hairier, but the logic is the same.

In the text:

<p>External commerce consists of the supply of the home market with foreign, and of foreign markets with home products.<a href="Endnotes.xhtml#fnCh1.1" id="ftCh1.1">[1]</a></p>

Endnote (in a completely separate file):

<p class="endnote">This is an example endnote.<a href="Chapter01.xhtml#ftCh1.1" id="fnCh1.1">[1]</a></p>

Now you have to specify which file the link points to:

Text: id = ftCh1.1, points to Endnotes.xhtml file + looks for fnCh1.1.
Endnote: id = fnCh1.1, points to Chapter01.xhtml file + looks for ftCh1.1.

Note: Calibre conversions create a hideous spaghetti nest of code.

First thing I would do is start renaming files from the unintelligible index_split_000.html -> Chapter01.html. That will make the code much easier to read. You'll then be able to more easily see what's pointing to what.

Originally Posted by Alcuin7

You built this with Sigil rather than Calibre? Making the footnote-in-footnote changes manually? Do I need to switch to Sigil? (I don’t want to do that if I can help it.)

Sigil or Calibre's Editor is perfectly fine. You can use either one for editing EPUBs.

(I personally just prefer Sigil 95% of the time.)

Originally Posted by Alcuin7

Since I know you are also familiar with MOBI, can I straightforwardly convert the EPUB (once I have it) into MOBI without further manipulation?

Yes, EPUB->MOBI works fine.

As long as the EPUB code works perfectly, you should be able to input the EPUB into KindleGen/Kindle Previewer, and get a working MOBI.

Originally Posted by Alcuin7

And how do I prevent MOBI from justifying the text, but keep it aligned left?

You could just not specify alignment in your CSS.

You can leave it up to the device/user to choose which alignment they like to read.

Note: Again, Calibre conversion creates a nest of "calibre##" CSS classes. You'll have to figure out which calibre## effect which parts of your book.

Originally Posted by Alcuin7

Are you saying that I need to use another format of endnote markers, say {a, b, c, …}, to identify my second-order notes, [...] This isn’t a problem, but if the tertiary (third level) are numerous – I haven’t counted how many there are – I could be running into four or five asterisks and daggers unless the symbols have their own specified order in Calibre.

I think thiago.eec hit the nail on the head.

You are saying there's 3 levels of footnotes, but I don't know if you're just numbering them 1-999 due to Word's limitations... or if your book would be better served with 3 distinct sets of footnotes!

Your book may be completely different, and it's hard to tell without seeing the DOCX firsthand.

Note: I was just specifying the code that my example PDF->EPUB used. 3 distinct sets of footnotes:

- Author
- Translator
- Editor

Nowadays, I went with a simpler: [1] + [T1] + [E1]. (You can see my 2018 remake here.)

For ebooks, I highly recommend against using the print-style symbols (*, †, ‡).

Alphabetical [a]-[z] is fine, but you have to watch out that you don't have more than 26 of a type of footnote in a chapter. (I once thought that would work for a book, but there was one chapter that had 50+, so [aa]-style becomes unwieldy.)

Your book might be different, but I highly recommend them all being in some sort of sequential order... and be human-understandable. :P

Even thiago.ecc's example of 31-a works fine for secondary footnotes.

Tertiary footnotes could potentially be 31-a-a... but at this point, I would probably recommend rethinking your document... since this becomes a hideous spaghetti nest and becomes too hard to follow/read. Books are ultimately meant to be read by humans. :P

Originally Posted by Alcuin7

You misunderstand. There is no way to embed footnotes into footnotes in Microsoft Word. It is accomplished by entering separate footnotes, formatting and cross-linking them so that they appear to be footnotes to footnotes, then hiding the real footnote markers.

If you're working on such a complicated document in the future, probably better to go to more advanced programs that can handle multiple sets of footnotes.

LaTeX has many packages that handle multiple sets of footnotes (bigfoot, manyfoot).

Originally Posted by Alcuin7

I assume I must also renumber “index_split_nnn.html” sequentially throughout the document so that they are in their correct order. [...] Is there some way to access them all? because if there is, then this becomes a fairly standard search-and-replace hack with external code.

The first thing I do is follow my Note above. Always rename the gibberish filenames to logical ones. It will make things SO MUCH easier to understand.

Then I also merge all files that need to be merged.

For example, your Endnotes file might be split into 100 "index_split_459" pieces. I merge them all into a single one and call it Endnotes.xhtml. Again, this makes the code so much easier to read/manipulate (and you can always split it in the future).

Originally Posted by thiago.eec

To see this in action, look at the file attached.

I would say a sample DOCX/EPUB speaks a thousand words. :P

[thiago.eec]

Originally Posted by Tex2002ans

Always rename the gibberish filenames to logical ones. It will make things SO MUCH easier to understand.

Then I also merge all files that need to be merged.

For example, your Endnotes file might be split into 100 "index_split_459" pieces. I merge them all into a single one and call it Endnotes.xhtml. Again, this makes the code so much easier to read/manipulate (and you can always split it in the future).

Exactly! I tried not to change the given code, so it would be clearer to understand, but the best practice is:

1. Polish the code (removing all those <sup> elements, for example);
2. Rename the files in a logical sense;
3. Merge all the notes (in a certain level) into a sigle file.

[Tex2002ans]

Originally Posted by thiago.eec

Polish the code (removing all those <sup> elements, for example);

Yep, and I also despise Calibre's <dd>-wrapped footnotes, so I instantly strip that code down to the barebones too.

IF you wanted to use EPUB3-style footnotes, you would have to use <aside> + some epub:type code. See Doitsu's Footnote test EPUB (and surrounding discussion about pop-up footnotes in that topic).

[Alcuin7]

Remember, I’m a newbie at electronic books. An old programmer (almost 40 years), but this is a new environment for me.

If I may press your patience, Tex2002ans and thiago.eec, I think I see several different ideas that I would like to gather together.

Footnote/Endnote identification

1. For first-level footnotes, I should use Arabic numbers.
2. For second-level footnotes, perhaps Roman numbers
3. For third-level footnotes, alphabetic is sufficient (there are not many)

Two-thirds of the book is endnotes/footnotes. There is a primary text with the normal short, regular, footnotes; many detailed explanations; and several side-excursions. Some of the detailed explanations and almost all the side-excursions also require footnotes. This explains the document’s complexity.

Many of the references come from Microsoft Word. <sup id="id_Ref489614211" class="text_6"> is one of these. I assume that I can dispense with these in EPUB and eliminate them to simplify my code?

Calibre converts the document into chapters, appendices, and a great many other divisions that make some sense. Many of the others make no sense. Am I required to follow Calibre’s divisions, or can I reassemble my own? Won’t that disrupt the order of the final product?

Where are these files index_split_000.html, index_split_001.html, …, stored? Do I merge/break/insert/rename within Calibre, or can I export them and deal with them outside Calibre? What is your normal procedure?


BTW, Tex2002ans, if you made the Say page for FEE, I think I used that writing my book, along with a great deal of material from Mises.org.

[Tex2002ans]

Originally Posted by Alcuin7

Footnote/Endnote identification

1. For first-level footnotes, I should use Arabic numbers.
2. For second-level footnotes, perhaps Roman numbers
3. For third-level footnotes, alphabetic is sufficient (there are not many)

Depends on the text.

Level 1 should be numbered footnotes: 1, 2, 3. I personally think this is the easiest to understand/read/reference.

Level 2 could potentially be: 1a, 1b, 1c.

Note: As an alternative to English letters, Greek letters could be used: α, β, γ, [...]. But I've only seen this in Physics/Maths, where people may be more familiar with the order of Greek characters.

Level 3... I HIGHLY recommend against this, this is 99% a sign your work is going to be too confusing. The information would be suited better:

• In the main text itself
• Shifted to another section (Appendix, Bibliography, [...])
• Moved into a normal footnote
• Removed completely

Originally Posted by Alcuin7

Two-thirds of the book is endnotes/footnotes.

... This sounds like a serious issue with organization of the material. Perhaps a lot more of that material can be put into supplemental Appendices, or new chapters.

And I know sidenotes... I'm extremely sidenote-/parentheses-happy. (Have you seen my posts on MobileRead?). But even I know where there's a limit!

When you start reaching notes about notes (Level 2), you really have to start relooking at making those points in your main text (or Level 1 footnote). Or just say: "More on this, see Jean-Baptiste Say, Chapter 3."

Think about it just like a Nested List. Once you start reaching 3+4 levels deep, probably time to think about a different way to present the info (Table, Chart, Graph, [...]).

Originally Posted by Alcuin7

Many of the references come from Microsoft Word. <sup id="id_Ref489614211" class="text_6"> is one of these. I assume that I can dispense with these in EPUB and eliminate them to simplify my code?

Well the Word-generated IDs could be important if anything needs to point BACK.

Personally, I strip EVERYTHING and make human-readable code.

Word-cruft:

<sup id="id_Ref489614211" class="text_6">1</sup>

Strip to Barebones:

<sup>1</sup>

Then use Regex to make Human-Readable:

<a id="ft1" href="#fn1">[1]</a>

If anything goes wrong, you can easily see what points to what. (And rely on fantastic tools like Calibre/Sigil's Links Report.)

With names like id_Ref489614211 + id_Ref489674211, your eyes will go cross-eyed.

Originally Posted by Alcuin7

Calibre converts the document into chapters, appendices, and a great many other divisions that make some sense. Many of the others make no sense.

By default, Calibre splits DOCX->EPUB files at:

• The 260KB mark
  
  • Old devices can't load large files. A single HTML file shouldn't be more than ~300 KBs, because old devices can crash.
• Every page break.
  
  • Your source DOCX may have these buried somewhere. (Manual page breaks after Title Page, page breaks before headings/images, etc., etc.)
• Every footnote.

These 3 settings can be tweaked in Calibre when you convert a book + go under EPUB Output.

Note: Under Structure Detection, Calibre splits files based on Xpath. So it'll look for common words like "Chapter/Book/Prologue" and split there.

Originally Posted by Alcuin7

Am I required to follow Calibre’s divisions, or can I reassemble my own?

You can reassemble them however you want.

Originally Posted by Alcuin7

Won’t that disrupt the order of the final product?

On the far left side of Sigil/Calibre, you see your list of your HTML files. That's the order EPUBs will read.

Originally Posted by Alcuin7

Do I merge/break/insert/rename within Calibre, or can I export them and deal with them outside Calibre? What is your normal procedure?

Best to do it within, Calibre/Sigil try to do a good job at updating/correcting Links.

So let's say you rename your:

Gibberish_split_001.xhtml -> Chapter01.xhtml

any links that pointed to Gibberish_split_001.xhtml will now point to Chapter01.xhtml.

You can also easily drag/drop files around, Right Click to Rename/Merge, etc. etc.

Originally Posted by Alcuin7

BTW, Tex2002ans, if you made the Say page for FEE, I think I used that writing my book, along with a great deal of material from Mises.org.

Nice! Yes, I do a ton of conversion work for them (have since 2013). If you've read an ebook from them, it was probably converted by me.

[Doitsu]

BTW, different apps/devices render footnotes differently. For example, iBooks requires footnotes to be properly defined, while the Kindle firmware (or KindleGen) uses heuristics and will usually display back-linked text as popup footnotes.

This works great for non-nested footnotes, however, if footnotes are nested, users usually can't tap links in footnote popup windows to display sub-footnotes.

I've created a simple epub3 test file with two different footnote sections.

The fist topic contains a link to non-nested footnote definitions:

Spoiler:

  <div class="footnotes">
    <div epub:type="footnote" id="fnid1" role="doc-footnote">
      <p class="endnote"><a href="../Text/test.xhtml#fnref1" title="footnote 1">[1]</a> Definition of the first footnote, with a link<a epub:type="noteref" href="../Text/endnotes.xhtml#fnid2" class="footnote" id="fnref2" role="doc-noteref">[2]</a> to the second foote. Donec metus ligula, elementum at pellentesque pellentesque, suscipit ac nunc.</p>
    </div>

    <div epub:type="footnote" id="fnid2" role="doc-footnote">
      <p class="endnote2"><a href="../Text/endnotes.xhtml#fnref2" title="footnote 2">[2]</a> Definition of the second footnote with a link<a epub:type="noteref" href="../Text/endnotes.xhtml#fnid3" class="footnote" id="fnref3" role="doc-noteref">[3]</a> to the third footnote. Donec metus ligula, elementum at pellentesque pellentesque, suscipit ac nunc.</p>
    </div>

    <div epub:type="footnote" id="fnid3" role="doc-footnote">
      <p class="endnote3"><a href="../Text/endnotes.xhtml#fnref3" title="footnote 3">[3]</a> Definition of the third footnote.Donec metus ligula, elementum at pellentesque pellentesque, suscipit ac nunc.</p>
    </div>
  </div>

while the second topic contains a link to nested footnote definitions:

Spoiler:

  <div class="footnotes">
    <div epub:type="footnote" id="fnid1" role="doc-footnote">
      <p class="endnote"><a href="../Text/test2.xhtml#fnref1" title="footnote 1">[1]</a> Definition of the first footnote, with a link<a epub:type="noteref" href="../Text/endnotes2.xhtml#fnid2" class="footnote" id="fnref2" role="doc-noteref">[2]</a> to the second foote. Morbi dictum luctus velit nec faucibus.</p>

      <div epub:type="footnote" id="fnid2" role="doc-footnote">
        <p class="endnote2"><a href="../Text/endnotes2.xhtml#fnref2" title="footnote 2">[2]</a> Definition of the second footnote with a link<a epub:type="noteref" href="../Text/endnotes2.xhtml#fnid3" class="footnote" id="fnref3" role="doc-noteref">[3]</a> to the third footnote. Morbi dictum luctus velit nec faucibus.</p>

        <div epub:type="footnote" id="fnid3" role="doc-footnote">
          <p class="endnote3"><a href="../Text/endnotes2.xhtml#fnref3" title="footnote 3">[3]</a> Definition of the third footnote. Morbi dictum luctus velit nec faucibus.</p>
        </div>
      </div>
    </div>
  </div>

(I used <div> instead of <aside> for backwards compatibility with epub2 readers.)

iBooks will display only the first footnote definition, if the definitions aren't nested. If they're nested, it'll display the footnote definition and all nested sub-footnote definitions, however tapping links in the popup window doesn't work. iBooks also seems to ignore styles in footnote popup windows.

eInk Kindles ignore nesting levels and will display only the first footnote definition. Users will need to tap Go To Footnotes to display sub-footnotes. However, if a sub-footnote link is tapped, its definition will be displayed in a popup window.

[thiago.eec]

Originally Posted by Doitsu

iBooks also seems to ignore styles in footnote popup windows.

iBooks does ignore the CSS classes, but it honors the style attribute within the element (at least some properties, anyway).

I use this code for my footnotes:

Spoiler:

HTML:

<aside id="footnote-1" class="Note-Text" epub:type="footnote" role="doc-footnote">
  <p style="text-align:justify"><a role="doc-backlink" class="Visible" hidden="hidden" href="Chapter1.xhtml#footnote-1-backlink">1</a> Text for the footnote. Text for the footnote. Text for the footnote.</p>
</aside>

CSS:

a.Visible {
  display: inline;
}

The footnote text will be justified on iBooks pop-up.

The "hack" using display:inline associated with hidden attribute is to prevent the number of the footnote appearing in duplicity on iBooks as well. iBooks generates a number title for each footnote, but since it doesn't accept classes inside the note, the css property will be ignored, while the hidden atribute will still be honored. This way, only the number on the pop-up title will be displayed.

The "hack" above does not affect other Reading Systems (at least not the ones I've tried), because the CSS class property (display:inline) overrides* the HTML atribute (hidden).

* See: http://www.hwg.org/resources/faqs/cs...tml#precedence

[Hitch]

Originally Posted by Tex2002ans

If you could attach a sample DOCX/EPUB, that would be helpful.

Agreed.

Yes, EPUB->MOBI works fine.

As long as the EPUB code works perfectly, you should be able to input the EPUB into KindleGen/Kindle Previewer, and get a working MOBI.

It should. As Doits pointed out (and thanks, Doits, so that I don't need to do that!) the popup footnote functionality is going to be flukey with 3 levels of "nested" footnotes.

You could just not specify alignment in your CSS.

You can leave it up to the device/user to choose which alignment they like to read.

Yes, and if you do set the alignment, e.g., "align left" you stand an excellent chance of getting whacked with a Kindle Quality Notice and ordered to "fix" it. Per the guidelines, you are not to set the alignment in your MOBI. (FWIW, I just slogged through an Earl Stanley Gardner Cool and Lam mystery with forced left-alignment, and OMG, it pissed me off to suffer through it. I absolutely reported it to the KDP, as the publisher is a big publisher and there's no excuse for it.)

I think thiago.eec hit the nail on the head.

You are saying there's 3 levels of footnotes, but I don't know if you're just numbering them 1-999 due to Word's limitations... or if your book would be better served with 3 distinct sets of footnotes!

Your book may be completely different, and it's hard to tell without seeing the DOCX firsthand.

This is why it would help to see the docx.

Hitch

[Tex2002ans]

Side Note: Long story short, Footception is not a good layout. See discussion about this on the LaTeX Stack Exchange:

https://tex.stackexchange.com/questi...hin-a-footnote
https://tex.stackexchange.com/questi...-footnote-text

Also, Word just isn't capable of doing this, you would need advanced tools that can generate multiple sets of distinct footnotes. Like the bigfoot package in LaTeX:

https://tex.stackexchange.com/questi...sted-footnotes

For book design/readability... it's almost always bad, bad, bad! And trying to hack this together in Word... atrocious, atrocious, atrocious!!!

Originally Posted by Hitch

Agreed.

Alcuin7 PMed me his DOCX, so I was able to take a look at it directly.

I stripped his DOCX down to a page and attached a sample DOCX:

AlcuinFootception[TexTest].docx

Note: ... and as expected, the numbering on his manually formatted "footnotes within footnotes" couldn't cope. ANY sort of adding/removing of footnotes is going to bamboozle the manual "level 2 footnotes".

* * *

From what I could tell, the footnotes were 1-999 because of Word's limitations to 1 level of footnotes.

His "invisible footnotes" were done using Word's Footnote functionality, similar to this:

Spoiler:

[...] the smallest²<sup class="tinywhite">3</sup> depositors.⁴<sup class="tinywhite">5</sup><sup class="tinywhite">6</sup> This is a perversion [...]

Note: I already explained this to Alcuin in PM, but these "invisible footnotes" are also an AWFUL solution for Search/Replace/Accessibility reasons.

So 3, 5, 6 appear at the bottom of the page in sequential order below the linebreak.

And then the footnote text below the linebreak:

Spoiler:

<hr />

² You can read the official agreement at [...] commence across Europe.¹¹⁷

¹¹⁷ This formation of the citation I owe to [...]

¹¹⁸ [...]

The bold 117 points to the "invisible" 117 (you can Ctrl+Click it in Word and jump to location).

Now, onto the bug. When you use Word's Save as HTML, this clickable link DOES NOT export to the HTML. It creates a simple ¹¹⁷ (same with Calibre's DOCX->EPUB).

So where you would EXPECT:

Spoiler:

<sup id="ft117" class="tinywhite">117</sup>

<hr />

¹¹⁶ You can read the official agreement at [...] commence across Europe.<a href="#ft117">¹¹⁷</a>

You just get:

Spoiler:

<sup id="ft117" class="tinywhite">117</sup>

<hr />

¹¹⁶ You can read the official agreement at [...] commence across Europe.¹¹⁷

I see a similar issue with his Cross Reference links within Footnotes too. The DOCX lets you Ctrl+Click like this:

Spoiler:

<hr />

¹¹⁶ [...] (See <a href="#ft73">footnote 73</a>)

but Word HTML export + Calibre's DOCX->EPUB turns into a plain:

Spoiler:

<hr />

¹¹⁶ [...] (See footnote 73)

URLs within footnotes look to be working fine, it's just these cross-links that don't appear in the HTML.

Very odd... anyone else more familiar with Word's Cross-Links functionality? Does HTML Export work for normal Cross-Links that are in normal text?

Note: And Alcuin's "cross-links" jump to the wrong location. They point to the "invisible" footnotes in the main text, NOT the footnotes themselves. So the HTML generated would be wrong anyway...

It should. As Doits pointed out (and thanks, Doits, so that I don't need to do that!) the popup footnote functionality is going to be flukey with 3 levels of "nested" footnotes.

Yeah, I'm highly recommending against nested footnotes for all sorts of reasons... ESPECIALLY when using Word. It just isn't built to handle this situation. And hacking something together is only going to bring extreme pain and misery.

Better to rewrite your document to include only a single level of footnotes.

Originally Posted by Hitch

Yes, and if you do set [...] forced left-alignment, and OMG, it pissed me off to suffer through it. I absolutely reported it to the KDP

Do they still not let the device manually override this?

It should be similar to something like Gitden Reader (or a Kobo). You have the Left + Justified buttons, but you also can just turn it OFF (Publisher Default).

User setting should override anything Left + Justified + without alignment set. And then shouldn't effect anything Center + Right aligned.

This is a common enough issue in ebooks, I would have assumed all newer devices (ESPECIALLY Kindles) should have this functionality.

It's just absolutely baffling to me.

[Hitch]

Originally Posted by Tex2002ans

Side Note: Long story short, Footception is not a good layout. See discussion about this on the LaTeX Stack Exchange:

Footception? Really? You're killing me here. :-)

For book design/readability... it's almost always bad, bad, bad! And trying to hack this together in Word... atrocious, atrocious, atrocious!!!

Well...in print, it would be relatively simple; you could indent the sub-notes. Perhaps that's the answer here? Altho....thinking about the "how" of how you'd do that. (Maybe indented bulleted notes?)

Alcuin7 PMed me his DOCX, so I was able to take a look at it directly.

I stripped his DOCX down to a page and attached a sample DOCX:

Attachment 167980

Note: ... and as expected, the numbering on his manually formatted "footnotes within footnotes" couldn't cope. ANY sort of adding/removing of footnotes is going to bamboozle the manual "level 2 footnotes".

We need Toxaris for this. I "know" that there's a huge issue surrounding Word's built-in x-refs, which is what you're talking about. The cross-referencing and indexing functions work in a way that's unique to Word, and you can't, for example, export it to another format and have it work in a different environment.

I forget now the "big experience" I had with this, but years ago, we struggled with a huge project that was a mess, over this very thing--that you couldn't just use the x-reffing capability, from point A to point B, product-wise. (like, Word to OO or whatever). I think it's that the x-reffing or index markers are done in a way that's not uniform or transportable. But honestly, Tex, I don't remember it that clearly and we should ask Tox. What he doesn't know about Word isn't worth knowing, in terms of the export to HTML. I'll go find him and ask him for some insight.

Note: I already explained this to Alcuin in PM, but these "invisible footnotes" are also an AWFUL solution for Search/Replace/Accessibility reasons.

Ah, good.

I see a similar issue with his Cross Reference links within Footnotes too. The DOCX lets you Ctrl+Click like this:



but Word HTML export + Calibre's DOCX->EPUB turns into a plain:



URLs within footnotes look to be working fine, it's just these cross-links that don't appear in the HTML.

And if you just export it to HTML, not using the Docx-->ePUB functionality in Calibre? Is it different, or no?

Yeah, I'm highly recommending against nested footnotes for all sorts of reasons... ESPECIALLY when using Word. It just isn't built to handle this situation. And hacking something together is only going to bring extreme pain and misery.

Better to rewrite your document to include only a single level of footnotes.

Yes.

Do they still not let the device manually override this?

As far as I know, the settings in the HTML still override pretty much everything. If the settings are not set, then the devices can override--but if you set align-left in the HTML, you force your choice on the reader.

It should be similar to something like Gitden Reader (or a Kobo). You have the Left + Justified buttons, but you also can just turn it OFF (Publisher Default)

.

That would be nice. Of course, you could also end up allowing users to override specific formatting (like blockquotes) too, and that would suck.

User setting should override anything Left + Justified + without alignment set. And then shouldn't effect anything Center + Right aligned.

LOL, there are probably people who would say that's wrong, too. You know how it is.

This is a common enough issue in ebooks, I would have assumed all newer devices (ESPECIALLY Kindles) should have this functionality.

It's just absolutely baffling to me.

I have the latest Kindle, AFAIK, the KFire 10, and I don't think it can override, even though it does allow user-selected line-heights/spacing and body text alignment.

Off to find Toxaris....

Hitch

[BetterRed]

Cross References in Word are implemented via Word Fields, the primary use of which is MailMerge and Excel integration. Word Fields can be nested. But I don't think foot/chapter/end notes use Word Fields, hence they don't inherit Field nestability.

BR