[Mister L]

Originally Posted by JSWolf

If you deactivate the CSS you get the default for <p> which leaves large paragraph spaces. You get no indents. If the section breaks use something visual, it's no longer centered and is now on the left. If italics and bold are created using spans, you lose those too. Smallcaps are also gone. In most cases, offset text will no longer be offset. And there are other issues. So please stop saying the eBook is made wrong. It's not made wrong. It's some of the defaults when there is no CSS that makes the eBook unreadable. CSS is there to override some of the defaults so the eBook is readable. The problems you've seen that caused you to deactivate the CSS are very easy to fix. No need to deactivate the CSS until they can be fixed. They can be fixed very easily. Just use the Calibre editor or Sigil and you'll be fixed in not long at all.

Most of those are aesthetic complaints, and as I said previously, I don't agree that an unattractive book is necessarily unreadable, it's just not as pleasant an experience. For some people it might be preferable (in particular cases where the CSS makes the book literally unreadable, such as in the case of low-contrast colours for colour-blind people). I personally am able and willing to edit the code to fix a book to my liking. Not everyone is, and they shouldn't have to. However the question of whether semantic code is more correct than unsemantic code is not a question of opinion; it simply is the case. As you say, italics and bold made using spans disappear if you disactivate the CSS; that's not the case if they are properly made with semantic tags, which has been more or less my point from the beginning.

Originally Posted by jackie_w

Unfortunately, yes. For general purpose epub readers it doesn't seem to matter whether you use <i>, <em> or <span class="italic">. All 3 may be visually OK (italic) but as far as TTS is concerned they may as well be plain text for all the difference it makes.

Okay, sorry, I misunderstood. That is a shame.

Originally Posted by jackie_w

BTW I'm not advocating against using semantically correct tags. In fact if I buy a book which uses <p> or <div> for headings I always change them to suitable <h?> tags as part of my clean-up workflow. But it's more a compulsion than an expectation of reaping any actual short-term benefits.

For your own personal use, if you aren't bothered by it, then sure, there is no reason to worry about the code (unless you are compulsive. I share that compulsion so no judgement ). For professionals it's a different story; I am capable of fixing a book if it was so poorly made that it distracts me while reading, but I get pretty grumpy about having to fix someone else's work on a book I've spent money for.

Originally Posted by Jim Chapman

Hi All,
Some of you may recognise me as the developer of the Freda e-reader app, which has been available on Windows for a very long time, and has been on Android for the past couple of years.

The thread interests me because I've lately been doing more work on:

• Properly parsing and applying CSS styles (using, believe it or not, my own hand-built xhtml renderer)
• EPUB3 compatibility (semantic markup, such as sections, and nav documents)
• Accessibility features (such as keyboard navigation by section and paragraph)
• Text-to-speech.

My new renderer comes with features (like Moon+ but perhaps a bit less daunting) to let you selectively turn of parts of the CSS styling (font selection, spacing, colour scheme). I would be very much interested to get feedback from the folks in this thread, regarding what Freda is getting right and wrong in this respect (and any other).

As regards the discussion on text-to-speech, so far, I didn't find any Android text-to-speech engines that give proper (or indeed any) treatment to the SSML emphasis tag. Whereas the Windows TTS engine does handle it reasonably sensibly.

Very interesting, I look forward to testing your app. It's refreshing to see a developer taking care to support CSS and accessibility features.