That said, it's also a terribly difficult format to work with, with extremely complicated code for the fixed positioning of everything. Which means that post-export surgery and fixup is nearly impossible. (Quite unlike r*fl*w*ble, which lends itself to all kinds of very elegant manipulation both pre- and post-export.)
All of which means: it's not really Kindle's fault, and no, there's not much you can do except try to optimize the export, with what little CSS adjustment is possible, and use it only for its one strength, picture-pages. No one, Adobe or other, is going to put further resources into "fixing" it at this late date. What little advancements there are are fixing more significant problems with reflow... well, you know.
But this doesn't explain why the same fixed-layout ePub file exported out of InDesign displays and works perfectly in Apple Books and in the old Kindle app (now called "Kindle Classic" v. 1.40.2). Fonts, formatting, styles, kerning, leading, etc., etc. all remain intact. Exactly. And all of the links within this particular book also work exactly as intended; ToC-to-chapters, linked text within the document, photos and graphics with links; doesn't matter if they're internal (to other locations within the book) or external (to outside sites).
I open the same exact purchased-from-Amazon copy of the ebook in the [old] Kindle app: Perfection. In a Kindle e-reader and the new app: Chaos. So, yes, it is Kindle's fault. And, to add insult to injury, they've taken a huge step backward with their new Kindle for Mac app, to boot. What did they change, and why would they do that??!!?
EPUB =/= PDF =/= Kindle. It's that simple. Three different formats, generated under three different sets of standards. And for Kindle, since there is no good way to directly generate their format, it's all up to how they choose to handle conversion from other formats.
As for then and now, the platforms and the readers are continually evolving. That one (so far) has evolved away from 'perfect' support of this largely obsolete format is unsurprising. Kindle has moved forward with a more capable format, overall, and has deprecated most of what came before, including its original MOBI format. It does not support EPUB directly, despite some marketing razzle-dazzle; it converts EPUB to its own current proprietary format whether it announces the fact or not. (Such as the supposed big deal of being able to side-load EPUBs; they aren't EPUB on the device.)
It's possible your book could be tweaked into better Kindle compliance, but again, one of FXL's weaknesses is that it is very difficult to work with once generated. If you're not familiar with the structure, reflowable EPUB is basically a packaged web site; a flowing text file with some markup that is presented, in liquid form, according to style rules. You can edit both the text and the styles pretty easily, if you have a little bit of HTML/CSS knowledge. FXL is...
I appreciate the time you've spent on this, and I'm sure your explanation makes sense. Unfortunately, I do not have the skill set to either verify or implement your suggestions. I am merely an "old-school" graphic designer who likes things to work correctly.
The from-InDesign ePub (fixed layout) book I'm referring to with all of this cannot even be opened by Kindle Previewer 3. Attempting to do so returns the error message [within the conversionLog]: "Kindle conversion has encountered an internal error while enabling Enhanced Typesetting on this book." Nevertheless, when published with KDP it displays and works entirely correctly in the [old] Kindle app.
"EPUB format can be read at least by the Kobo eReader, Blackberry Playbook, Apple's iBooks app running on iOS devices such as the iPhone, iPod Touch and iPad, Barnes and Noble Nook, Sony Reader, BeBook, Bookeen Cybook Gen3 (with firmware v. 2 and up), COOL-ER, Adobe Digital Editions, Lexcycle Stanza, BookGlutton, AZARDI, FBReader, Aldiko, Moon+ Reader and WordPlayer on Android, Freda on Windows Mobile and Windows Phone 7, the Mozilla Firefox add-on EPUBReader, and Okular. Several other desktop reader software programs are currently implementing support for the format, such as dotReader, Mobipocket, uBook.
"Kindle devices do not support the EPUB file format used by many other e-book readers. Instead, they are designed to use Amazon's own e-book formats: AZW, MOBI and, in later devices, KF8. These formats are intended for reflowable, richly formatted e-book content and support DRM restrictions, but unlike EPUB, they are proprietary formats.
"It is unlikely Amazon will support EPUB3 (they could have done that with KF8). Amazon has valid reasons for using KF8: it has a built-in dictionary index and fixed layout, which EPUB does not really have yet, (but currently under development)."
The first of two crucial takeaways here is that hardly any two EPUB readers manage the format and its presentation the same way. There is a standard, frequently updated (these days, largely with nits about accessibility, with no attention paid to major faults and needs for upgrade), but there is not even a weak enforcing body or one with enough "gravity" to pull reader developers into a common orbit. That your exports work on Apple is very mildly surprising; while their reader is fairly standard, Apple follows its general principle of doing things "differently and better." They at least have the engineering muscle to to things right as far as code quality and product testing. Many of the other readers come from very small shops, if not solo developers, and incorporate every quirk the maker can think of, often with minimal stability and compatibility testing.
The other takeaway is that the production tools only take the job so far. There is no perfect EPUB generator. It's not like using InDesign to PDF to a print service, where zero outside (or "internal") knowledge is needed; just use the app right and all will shine. But with a fragile, mishandled format like EPUB, InDesign (for example) does its best, but that 'best' has limits not addressable from within the app, and a truly superior result requires skills and knowledge outside the menu-driven interface.
Yes, I'm also "old-school" enough to have done wax pasteups and drawing-table drafting. And to have been in the midst of the conversions to Postscript and PDF, and the adaptation required for same using QuarkXPress and InDesign sice, almost, Day One. Also a pre-release beta tester for Photoshop, after having used Scitex, Full Color Publisher, and Letraset ColorStudio.
But coding; damn! For whatever reason, the old grey cells balk. It's not that I don't want to learn new skills, but part of the current situation is that I simply don't have the time to do it quickly enough to solve the ongoing problem(s). On the to-do list... but not very near the top.
What can be done is to export to "Adobe PDF (Print)" (PDF/X-1a format recommended) from InDesign, and then run the PDF through Kindle Create. The problem is that none of the PDF/X formats allow hyperlinks. So I tried doing this by exporting as "Adobe PDF (Interactive)" and it worked, more or less... but Create also broke about half of the links (including all of the links in photos), and messed up the formatting of the text links that it didn't break. So... not a complete solution.
I'm trying to decide whether to buy a Kindle 2, now that it's available in Europe. One thing I would like to do with it is to put several Windows help files on it (for example, the Python documentation). Is there a good way to convert a CHM file to a Kindle-readable format?
There are commercial programs like ABC Amber CHM Converter that will perform steps 1 & 2 automatically. However, I don't know any software that will convert CHM directly to EPUB. Other commercial applications (for example, Universal Document Converter) can be used to convert CHM to PDF which the Kindle may or may not handle all that well.
In order to read your .chm file on the Kindle, you will need to convert it to .mobi format. That can be done by merely using Calibre which allows you to import a .chm file and convert it to .mobi. Then you can send the file to your device when it is connect
I have found Sigil to be an excellent tool for creating and editing ePub format e-Books. I use it on Linux, but as far as I know it works equally well on Windows or Mac, so it should not matter at all what operating system you are trying to accomplish this task on. As long as you have the HTML and image files you need (and know enough about HTML to possibly edit the image links or clean up the content if necessary) you should be fine. It requires a bit more work than eCub, but it gives you total control over the end result and with a little practice you can create some quite nice ePub e-Books.
While using Publisher to format the printed version, I needed to tweak the text so it fitted the physical space available on particular pages. That's one reason why the text in Publisher is different to its original sources. But as far as I can see, the only way to extract text from Publisher at present is via copy / paste?
Sorry, I don't have enough experience in that area to say. If no one here knows, and you investigate that approach further, you might check with the developers of whatever tool you are considering to get their advice on that.
Producing well-formatted eBooks is rather like the problem of producing webpages when all web browsers were different and worked to different standards: With the ePub output from PagePlus, I have used a code editor to inspect the output coding, line by line, in order to understand & overcome problems: A badly encoded eBook may well work OK with some eReaders & eReading software, but fail badly with others.
c80f0f1006