How does one join epub files?

[BobC]

I had a look at the Wikisource version and I see it shows that it is derived from a version at Gutenberg.

This could have accounted for how the Wikisource version has been split, though I see Wikisource has it split amongst 23 files whereas Gutenberg has managed it across 10 (but then the automated approach by Gutenberg does seem to split without reference to chapters etc and probably simply by size).

There is no reason to assume the Wikisource version is split according to original publication, however looking at some of the files they do seem to split logically (no obvious split in the middle of a conversation for instance) and would make a good guide to where to split.

As mentioned earlier in the thread the Adelaide version doesn't have the italics that can be found in the Wikisource (or Gutenberg) versions. You takes your choice as to which is the best to use as a basis, they all have pros and cons. Gutenberg would probably need all the files merging into one and then re-splitting but then you get the italics, Adelaide needs splitting but there are no italics and with Wikisource you need to download each section separately and then build them into a single EPUB. To what degree both Wikisource and Adelaide have been proofed and corrected against a paper copy isn't clear, so how far they diverge from the Gutenberg text is something that might be worth looking at before choosing.


BobC

[theducks]

I agree, it seems the device is doing this.
Odd, it is not near the 260K (safe) break point needed.

IMHO, it might be better to find a Story logical (scene break) and split the section into 2 parts there.

[AlexBell]

Originally Posted by Tex2002ans

And Alex, like JSWolf said, once you split those files into separate HTML files, there just isn't much you can do. Each HTML file starts on its own page, you don't have control over that. It just so happens to be one of the quirks of reading systems we have to live with.

Thanks for all your other comments. But I'm afraid I don't agree with the above part of your response. There are no chapters in Cousin Pons. The Project Gutenberg version opens with the first line, and one can read to the end of the book without starting a new file. I've opened it up, and there are several HTML files within it. And Crutledge's version of A Distinguished Provincial in Paris in the MR library also does not have any chapters, and does include several HTML files containing successive parts of the story. But the text as read on my Sony is continuous, as it was in Balzac's original book.

I think the answer is in the content.opf file, and am trying to decipher the structure of Crutledge's ebook. But it was done in Sigil, (which I don't use) so it's hard to know what makes the difference.

And thanks to everyone else who responded to my post yesterday. So far as italics and diacritics are concerned I think it's part of my job to check with the original, and insert them where necessary.

[Tex2002ans]

Could you link over to the files? It's hard to make judgements on this stuff without seeing the actual code.

[BobC]

Originally Posted by AlexBell

Thanks for all your other comments. But I'm afraid I don't agree with the above part of your response. There are no chapters in Cousin Pons. The Project Gutenberg version opens with the first line, and one can read to the end of the book without starting a new file. I've opened it up, and there are several HTML files within it.

What are you using to read the Gutenberg version with ? I have a copy and whatever viewer I use I get a new display page when I move between individual HTML files within the EPUB.

So, for instance

Well, I myself believe that there is an intelligence in works of art; they know art-lovers, they call to them—‘Cht-tt!’”

which is at the end of the first html file (OEBPS/@public@vhost@g@gutenberg@html@files@1856@1856-h@1856-h-0.htm.html) is on one page then

Mme. de Marville shrugged her shoulders and looked at her daughter; Pons did not notice the rapid pantomime.

which is at the beginning of the next HTML file (OEBPS/@public@vhost@g@gutenberg@html@files@1856@1856-h@1856-h-1.htm.html)starts on a new display page.

Of course it is possible that by happenstance depending on the font used and the other display settings on your device the end of one HTML file corresponds with the end of a display page, giving the illusion of continuity between the separate HTML files that make up the EPUB, however what is certain is that each HTML file will start to be displayed on a new page - that is the nature of the viewer and nothing to do with the OPF which will, amongst other things control the order in which the HTML files follow each other for display..

Are you aware that there is a copy of Cousin Pons at Archive.org - this is in the form of an image of the book and as you say it is just one long text without any obvious scene changes or chapters. I've used the DJVU versions of Archive.org texts previously when proofreading; the hidden text layer can be useful for locating where certain words occurs so you can examine the original image of the page.

BobC

[AlexBell]

Thanks Bob.

I use a Sony PRT3. I'll go back again and check, and get back to you.

Yes, you were right and I was wrong. I'll have to crawl back into my little hole and start over.

[BobC]

The "new file = new page" is an intrinsic part of EPUB. If you have ever played with a monolithic format such as FB2 where the book is just one long file you would understand why large files are a problem and EPUB breaks them up into effectively segments that a small handheld device can handle.

With FB2 and a long book such as the King James Bible or Seven Pillars of Wisdom it becomes increasingly slow to progress through reading the book and even opening it in the first place can either be glacially slow or even impossible.

While it might be possible to deal with a mega-sized file in FB2 or a single-segment EPUB on a desktop computer with Gigabytes of RAM and the ability to use a swapfile, in reality that's not where you would want to read most Ebooks.

The Terry Pratchett "Discworld" books tend to be, just like "Cousin Pons" unbroken by chapters and too large for a single file. These commercially produced books also have to artificially decide where to split the files with the consequent new page.

As @theducks mentioned before it's a matter of where to split your file and into how many parts, personally I'd want a minimum of four to ensure maximum speed/performance; it's a matter of locating suitable places to do the splitting. The Wikisource version might give some hints with where they have chosen.

BobC

[Sarmat89]

Originally Posted by BobC

If you have ever played with a monolithic format such as FB2 where the book is just one long file you would understand why large files are a problem

If you use an event-based parser, you don't need excessive memory. All you need is to scan the file once, and build the lookup table. Formats actually optimized for portative devices, like MOBI or PDB, have those tables pre-built.
The actual reason EPUB needs segmentation, is a bad container format which doesn't allow random access, and the atrociously complex HTML+CSS+JS content format, which needs a DOM tree created just to make heads and tails of it.

[AlexBell]

Thanks again.

I've decided that I am going to learn how to design ePub3 ebooks, if I can, before ePub4 rears its head.

I have 'EPUB 3 Best Practices' by Matt Garrish and Markus Gylling on order, but it will take 2-3 weeks to get to Launceston. Can anyone recommend any other teaching resources for learning ePub3?

When I began to design ePub2 ebooks I made myself templates for chapter files and config.sys and toc.ncx files. I'd like to do the same for ePub3 ebooks if I can. Can anyone recommend a public domain ePub3 ebook, perhaps in the MR library, that I can download and open up to see what the files look like?

[Doitsu]

Originally Posted by AlexBell

Can anyone recommend any other teaching resources for learning ePub3?

You could simply convert one of your epub2 books to an epub3 book with Sigil and the epub3 output plugin.

Originally Posted by AlexBell

Can anyone recommend a public domain ePub3 ebook, perhaps in the MR library, that I can download and open up to see what the files look like?

O'Reilly offers some free epub3 books for download. For example:

The Little Book of HTML/CSS Coding Guidelines

The IDPF also offers several epub3 sample books for download:

EPUB 3 Samples Project

[AlexBell]

Originally Posted by Doitsu

You could simply convert one of your epub2 books to an epub3 book with Sigil and the epub3 output plugin.


O'Reilly offers some free epub3 books for download. For example:

The Little Book of HTML/CSS Coding Guidelines

The IDPF also offers several epub3 sample books for download:

EPUB 3 Samples Project

Thanks, that's most helpful.

I'm afraid I'm old and set in my ways, so I'd rather continue designing my ebooks myself rather than have Sigil do it for me. I'll download the IDPF sample books and open them up.