Pratchett and Chapters?

[enuddleyarbl]

This isn't really an ePub question, but I don't really know where else to ask it. I'm editing my copies of Pratchett's Discworld books. Apparently, Pratchett doesn't like chapters. Essentially, the book is one big blob of text with some classes that work out to thematic breaks (which I have replaced with <hr/>s). That blob of text has been arbitrarily broken into two or three files (either by the publisher or by Calibre). But, when I read the book on my ereader (Kobo Forma), there's a noticeable delay (3-5 seconds) when I hit those loading points.

I was considering changing the <hr/>s to <h3>s (everything else is <h2>) and having Calibre's editor automatically split the book at those points. But, Pratchett's books tend to have someplace around 70 thematic breaks throughout them. So, that means a silly amount of "chapters" (i.e., files). Is that something reasonable to do? Is there some limit to the number of files in a book? I guess since the <h3>s won't have any text associated with them, I don't have to put them into the ToC. But, is there a better alternative?

[Tex2002ans]

Originally Posted by DaveLessnau

I'm editing my copies of Pratchett's Discworld books.

I'm in the middle of reading Sourcery right now.

Originally Posted by DaveLessnau

Essentially, the book is one big blob of text [...] That blob of text has been arbitrarily broken into two or three files (either by the publisher or by Calibre).

In Sigil or Calibre, go into the Reports:

• Tools > Reports (Ctrl+Shift+R)

Go into the "Files" Reports and sort by filesize.

It's good practice to keep each HTML file <300ish KBs. (This was a limit on really early, really old ereaders with very limited CPU/RAM.)

What Should You Do If The Files Are Too Large?

Split them up.

The easiest way I do this is:

• Merge everything into one monolithic file.
• Skim through the merged file, placing "split markers" every so often.
  
  • In your case, before scenebreaks.
• Split the book.

I explained the Sigil way in step-by-step instructions here:

• 2021: "Can't open epub file"
• 2021: "Advice for fixing a publisher's broken book"

Calibre's way is very similar.

- - - - -

Side Note: All you have to do is make sure you don't have HTML files >300 KBs. The devices should be able to work fine from there.

Sometimes, if the book is in one huge, multi-MB HTML file, things can be pretty sluggish.

Originally Posted by DaveLessnau

But, Pratchett's books tend to have someplace around 70 thematic breaks throughout them. So, that means a silly amount of "chapters" (i.e., files).

So?

Some books have chapters that are a single page. Or a single word. So what?

Originally Posted by DaveLessnau

Is there some limit to the number of files in a book?

Heh.

There are extreme cases where the OPF or NCX file has reached >300 KBs.

In all these years, I don't think I've really heard much about it though.

You won't be reaching that any time soon, unless you were trying to merge decades and decades of articles together into one monolithic omnibus. :P

Side Note: In all the ebooks I've done, there's maybe a handful I ever ran across where a single chapter was >300KBs.

Originally Posted by DaveLessnau

I guess since the <h3>s won't have any text associated with them, I don't have to put them into the ToC. But, is there a better alternative?

I find Sigil's splitting to be much easier with the simple:

<hr class="sigil_split_marker" />

and Edit > Split at Markers.

... but in Calibre, you can also split based on anything—like your custom <hr>s—by using XPaths.

Just:

• Right-Click an XHTML file > Split at Multiple Locations

You can read a little bit about it here:

• Calibre Manual: "Splitting HTML Files"
• Calibre Manual: "XPath Tutorial"

[Jellby]

There can be thousands of files in an ePub (e.g. 1001 Nights). It may cause some apps to slow down or use more resources (or maybe the opposite), but no Pratchett book will be that extreme.

However, an undesired effect of splitting the book in chapters is that you'll (probably) get a page breaks at every chapter. I'm not sure Pratchett would have liked that (perhaps it was only the publisher who decided not to have page breaks in order to save paper). So at the end I decided to ignore filesizes and put the whole book's text in a single "Text.xhtml" file. At least KOReader handles these large files alright.

One thing you may have noticed (depending on which "edition" you're reading) is that some scene breaks are just plain spacing, while others have a star, a row of asterisks or something else. This is just a direct translation of the print version without any understanding of the meaning. There is no (semantic) difference between these two types of scene breaks. In print they're all initially just spacing, but when they fall exactly at a page boundary, an asterisk is printed to make it explicit, as otherwise it would be almost invisible. In an ebook, there's no way (without javascript) of having this effect, so I opted for having an asterisk at every scene break.

[JSWolf]

What I do is fix the Discworld books so each break breaks at a section break. I don't find it as slow to load the next html file. But then, I make sure I'm not making them too large when I do this.

[enuddleyarbl]

@Tex2002ans: I keep forgetting the Reports tool in Calibre's editor. I checked on the current book I'm reading and the two "text" files are about 200KB each. So, it shouldn't be a problem. But, the loading is noticeable. I'd also forgotten that I could tell Calibre to split on other things than headings (h1, h2...). Thanks.

@Jellby: I was wondering about the different types of breaks I found in the original book. Their trying to keep the book looking the same as the print version explains it. Thanks.

@JSWolf: I don't recall seeing actual section breaks (<section>) in the books. Are you including things like <hr/> in there. I'm assuming you added them manually. How do you determine what's a good size for a break?

[JSWolf]

Originally Posted by DaveLessnau

@JSWolf: I don't recall seeing actual section breaks (<section>) in the books. Are you including things like <hr/> in there. I'm assuming you added them manually. How do you determine what's a good size for a break?

You won't see <section>. Take the eBook Thief of Time. A section break is

.paraNoIndent	{
display: block; 
margin-bottom: 0em;	
margin-top: 1em;
text-indent:0em;
}

<p class="paraNoIndent">

As for section breaks, I use...

hr {
  margin-top: 0.9em;
  margin-right: 40%;
  margin-bottom: 0.9em;
  margin-left: 40%;
  border-top: 2px solid;
}

[enuddleyarbl]

@JSWolf: In the previous Pratchett books I'd edited, there was some variation (or two (or three)) of a section break class as you described. And, like you, I'd replaced it with my <hr/>. In the current book I'm editing, they didn't use such a class. It looks like they're using a pairing of empty paragraph marks and their first paragraph in the chapter class:

<p></p>

<p class="chapteropenertext">

Good thing I looked. In the previous books, with only one "chapter," there was only the initial use of that <chapteropoenertext> tag. I would have routinely deleted it (Replace All w/ nothing) if I hadn't been looking for the breaks you mentioned. Thanks.

[JSWolf]

Originally Posted by DaveLessnau

In the current book I'm editing

Which book is this? I can have a look and see what the code looks like. It could be that you have an old version and the code has since changed.

[enuddleyarbl]

It's "Wyrd Sisters" (Discworld #6). There's no need to look at the code. I've looked it over and just replaced those groups with <hr/>. There were also 8 situations that had an anchor (<a...>) instead of the paragraph mark pairs. And I replaced those as well.

[JSWolf]

Originally Posted by DaveLessnau

It's "Wyrd Sisters" (Discworld #6). There's no need to look at the code. I've looked it over and just replaced those groups with <hr/>. There were also 8 situations that had an anchor (<a...>) instead of the paragraph mark pairs. And I replaced those as well.

Current code for section breaks for Wyrd Sisters is

<p class="para">*</p>
<p class="paraNoIndent">

I don't need regex to search/replace the code.

[enuddleyarbl]

Odd. I thought I'd looked for the section breaks before doing any editing. It's possible I deleted the para tags already. But, I'm almost positive there wasn't a ParaNoIndent in my copy. Oh, well. I managed to find them and replace them regardless.

Also, I had the Calibre editor break the files at those <hr/>s. In this book, there are about 90 book text files now. It all seems to work fine and the delay upon loading is gone. Thanks.

[JSWolf]

Originally Posted by DaveLessnau

Odd. I thought I'd looked for the section breaks before doing any editing. It's possible I deleted the para tags already. But, I'm almost positive there wasn't a ParaNoIndent in my copy. Oh, well. I managed to find them and replace them regardless.

Also, I had the Calibre editor break the files at those <hr/>s. In this book, there are about 90 book text files now. It all seems to work fine and the delay upon loading is gone. Thanks.

I don't break the HTML files at every section break. I do it such that when the HTML has to be broken, I do so at a section break.

[Quoth]

Some of his books just have anonymous chapter breaks.
Dune has epigrams (extracts from a fictional history book) and no page breaks or chapter headings.

I don't think ebooks should slavishly follow paper books, so if the scene break is really an anonymous chapter start I'd be happy with a page break. Similarly with Dune. I have a lot of Dune on paper and recently got the 1st 3 cheap as ebook, so I will put page breaks and TOC to the start of each "history" quote.

I have all Discworld and much other Pratchett on paper, so unless it's legitimately offered very cheap for the series…
Or I discover I'm living for 100s of years.

[Tex2002ans]

Originally Posted by Jellby

One thing you may have noticed (depending on which "edition" you're reading) is that some scene breaks are just plain spacing, while others have a star, a row of asterisks or something else. This is just a direct translation of the print version without any understanding of the meaning.

This is why you should always put asterisks or SOMETHING as scenebreaks. Not just rely on a gap.

Print book ≠ Ebook.

See my:

• 2019: "Why is it so hard to preserve blank lines?"
  
  • Especially the EPUB Secrets article I quoted/linked to.

Originally Posted by Jellby

There is no (semantic) difference between these two types of scene breaks.

Indeed.

In this specific case of Discworld books, there is no difference in scenebreaks (I'd trust Jellby's judgement on that).

Depending on the book, you have to go through and make sure.

(But in 99.9% of all books you encounter, they're normal and don't have multiple types of scenebreaks. )

Originally Posted by Jellby

In print they're all initially just spacing, but when they fall exactly at a page boundary, an asterisk is printed to make it explicit, as otherwise it would be almost invisible.

In an ebook, there's no way (without javascript) of having this effect, so I opted for having an asterisk at every scene break.

Yep. Exactly. And I agree with your solution completely.

- - -

Side Note: If you want more details, this scenebreaks discussion already occurred in:

• 2020: "Problem adding extra lines in text"
  
  • Especially my Post #25.

We covered lots of examples + all the technical details you'd ever need.

(Even bleeding-edge Javascript to try to emulate "only show scenebreaks at the bottom of pages/screens"... but I definitely wouldn't choose that "solution".)

- - -

Side Note #2: If you wanted even more extreme technical details, I described some of the "surgical" approach to editing in:

• 2021: "Indesign-epub-kindle formatting problem: footnotes export with massive indent."

but this is really getting into the weeds.

This explains how/why you may not want to wipe everything away, and clean/edit the ebooks with more precise cleanup methods.

Again, 99% of the time, you're working from complete garbage, so it's easier to wipe everything and start from scratch.

But in some cases, markup might be left in the original ebook—like "invisible" scenebreaks—which may make your life easier when trying to fix EPUB->EPUB.

Originally Posted by DaveLessnau

@Tex2002ans: I keep forgetting the Reports tool in Calibre's editor.

The Reports are AMAZING. Definitely take lots of advantage of those.

You can also use them to list all links in a book, making it very easy to:

• Spot broken footnotes
• Visit every single URL and let you know it's alive
• Or lots of other tricks.

Originally Posted by DaveLessnau

I checked on the current book I'm reading and the two "text" files are about 200KB each. So, it shouldn't be a problem. But, the loading is noticeable.

Originally Posted by DaveLessnau

Odd. I thought I'd looked for the section breaks before doing any editing. It's possible I deleted the para tags already. But, I'm almost positive there wasn't a ParaNoIndent in my copy. Oh, well. I managed to find them and replace them regardless.

Sometimes the person who originally digitized the book made a mistake as well. This is why, if you're extra serious, you try to get a copy of the physical (or PDF), so you can figure out what the heck is going on.

Like I said, 99% of the time, it's probably just your typical one-type-of-scenebreak... but there are very odd things out there.

Note: Like the Wild Cards books I read many years ago, they used a ♥, ♦, ♣, or ♠ for certain scenebreaks.

Originally Posted by DaveLessnau

Also, I had the Calibre editor break the files at those <hr/>s. In this book, there are about 90 book text files now. It all seems to work fine and the delay upon loading is gone. Thanks.

Did you have complicated CSS?

The Kobo definitely shouldn't be chugging on a file like that.

[JSWolf]

Part of the problem with Discworld is that it uses a 1em blank space for the spacebreak. When I was using blank space for a section break, I used 2em. 1em just doesn't work.