Editio Self-Publishing discussion

Hi:

After working with the lead developer of Sigil, I have managed to create a commercially compliant ePub using nothing but Word and Sigil. No need to use a third-party software like Calibre.

I have created a free circular with detailed directions. It can be downloaded from my website.

Thanks, that's very useful. I looked at Sigil a while back and might do so again now. I didn't realise you could create the epub structure so easily from the html export from Word.

I don't disagree with anything you've written, but I would add to anyone reading that:

* Be sure to follow Michelle's advice to split files at chapter breaks. This seems to encourage page breaks on Adobe Digital Editions and also if you want a Kindle version and have a top margin on your chapter headings, this is the only way to achieve that for most Kindle configurations.

* Getting a Kindle version... If you take the ePUB output and run KindleGen over the epub file (which you can do most simply by opening the epub file in the Kindle Previewer from Amazon) then you should get a good quality Kindle version. However, several current Kindle reader apps, and the Kindle 3 running firmware older than 3.4 don't understand html elements with more than one class, which is a problem I found all over the place when I tried using Calibre. You don't need to know html to see if you have a problem. If you have paragraphs that are all italic, bold, centred or indented differently from standard text, just view your book in the Kindle Previewer using the 'Kindle' device setting, and verify those paragraphs have come through okay.

Am I assuming correctly that this is the problem with the iPad and MacOS apps for Kindle also?

Tim: there are two problems with going from ePub to Kindle. The first is that ePubs only create a metadata TOC while Kindle books require an HTML TOC. Many Kindle books -- those made using InDesign for example -- have both a metadata TOC and an HTML TOC, but the "Go to Table of Contents" option in Kindle only points to the HTML TOC. Without this HTML TOC, the option is greyed out.

The second problem is that, while Amazon will tell you they accept ePubs for conversion, dig a little deeper and they will tell you they prefer the HTML be ONE file, not split as is the case with ePubs.

The only way around these problems going from ePub to Kindle is to build a table of contents in Word and leave it there. Then you build another one in Sigil. Then BEFORE you split the HTML document into separate files for ePub you save the unsplit ePub under a new name and use that for conversion to Kindle.

For your ePub you can then either remove the HTML TOC and split the document up as required OR you can keep the HTML TOC in and have the benefit of both TOCs in your ePub BUT you will then have to fix all the HTML TOC hyperlinks to point to the new HTML files created after the split. Otherwise you will end up with an HTML TOC whose links will all simply take you to the start of the same chapter.

Hi David,
Yes, I think so. If I sound vague, it's because I don't have a Mac to prove what you say. But some of the device options (notably the iPAD) on the Kindle previewer work the same way as older Kindle devices, only reading the first class for each entity (and so losing some formatting). Trusting the Kindle previewer is a little dangerous because it doesn't always match what happens on a device (especially since the device will have further firmware and o/s combinations) but I've heard this confirmed from iPad users that the reader for iPad does this too.

M.A. wrote: "Tim: there are two problems with going from ePub to Kindle. The first is that ePubs only create a metadata TOC while Kindle books require an HTML TOC. Many Kindle books -- those made using InDesign..."

Thanks for all your valuable information.

Yes, very good point about the HTML TOC because that would need to cope with any file splitting. I'm glad you raised it. About half the books I build are novels, so I'm not so bothered about the HTML TOC. It is only the logical TOC (the ncx file) that Amazon state is mandatory in their coding guidelines.

I'm not sure where Amazon are going with their TOCs, but with the most recent Kindle device I have (Kindle Touch) will show either the hierarchical logical TOC directly in the Goto menu or (if there is no logical TOC) will show the old 'Table of contents' entry that jumps to the HTML TOC. Hopefully the HTML TOC will soon die, to leave behind the same kind of contents pages you get in printed layouts, such as a list of stories in an anthology.

Thanks for your help

Tim: you still need to be bothered about the HTML TOC. In the Kindle Fire, if a mobi file only contains an NCX TOC the option Go to Table of Contents is greyed out and no option exists to access the NCX file. Ditto for Kindle for PC. So why Amazon make the NCX and not the HTML TOC mandatory is ass-backwards. I can only imagine it is because they will eventually migrate toward using the NCX file. But in the meantime readers have no TOC with which to navigate easily around the ebook.

I think you are mixing up the HTML TOC and NCX TOC. It is the HTML TOC that creates a contents page similar to what you find in a print book, where you can lay it out after your title page and the reader can scroll past it and into the book like one does with a print book. The NCX TOC is the opposite: it operates as a separate file that you cannot scroll out of to enter the book; you must click on a link to get back into the book file.

Try it and you'll see what I mean.

M.A. wrote: "Tim: you still need to be bothered about the HTML TOC. In the Kindle Fire, if a mobi file only contains an NCX TOC the option Go to Table of Contents is greyed out and no option exists to access th..."

Thanks, as always, for the advice.
Tim

David wrote: "Am I assuming correctly that this is the problem with the iPad and MacOS apps for Kindle also?"

Hi David, I haven't tried it yet, but I've just seen the spec for Kindlegen 2.7 According to the spec, it fixes this problem by building Mobi7 format files that permit more than one class for each element. Using Kindlegen to convert ePUB to Kindle format suddenly looks a much safer proposition.