CSS ignored in .mobi (on Kindle)

[_savage]

Hi,

I've spent the last few days putting together an eBook in .epub format which renders just fine on my Nook. In order to make it available for Kindle I used this website and it seems to do a decent enough job. (Turns out that the latest kindlegen (V2.4 build 69900, link) completely breaks the encoding of some of the Pali letters, and also bloats the final .mobi file to twice the size of the original .epub file, even with -c2 compression.)

However In my .epub I used HTML quote tags like so

<q class="bla">This is a quote.</q>

and I style them as follows

q {
  quotes: none;
}
q.bla {
  font-style: italic;
}

I explicitly do not want to have quotation marks of any sort around the quote! While this works fine on my Nook and in a web browser, it seems that a 2 year old Kindle still puts quotation marks around these quotes.

Another issue are unordered bullet lists. Again, I style them to disable the bullets altogether

ul.toc {
  list-style: none;
}

only to find that my table of content

<ul class="toc">
  <li> <a href="foreword.html">Foreword</a></li>
  ...
</ul>

does render with bullets on that same Kindle.

How can I get rid of both, the quotation marks and the bullets; or is this a Kindle specific issue?

Thanks!

[susan_cassidy]

.mobi does not directly support css. It can only translate some css into equivalent plain HTML, so these constructs will not work for .mobi.

[Jellby]

And it is the converter which does the translation. Some converter may support some CSS, another converter will support a different set of CSS, and some things may be difficult/impossible to translate to mobi.

[Oldpilot]

Originally Posted by Jellby

And it is the converter which does the translation. Some converter may support some CSS, another converter will support a different set of CSS, and some things may be difficult/impossible to translate to mobi.

What does that mean? The ultimate "converter" is Kindlegen, surely?

[Oldpilot]

Wouldn't it be easier to go <i>This is a quote</i>?

I seem to recall that the Publishing Guidelines say (or used to say) that one can't use a table in the "table" of contents.

http://kindlegen.s3.amazonaws.com/Am...Guidelines.pdf

[DiapDealer]

You might want to also note that the <q> tag is only supported in the KF8 format (and only on the Kindle Fire) ... not the regular MOBI format.

The Amazon Kindle Publishing Guidlines gives you examples of several methods that can be used to specify separate CSS statements to tailor the MOBI portion in relation to the KF8 portion (and vice-versa). In short: starting with a validating ePub does not mean kindlegen will build from it, a MOBI/KF8 version that will be identical to the original ePub in every way. You have to learn what input Kindlegen will play nice with. The Amazon Kindle Publishing Guidelines is the place to start

Also, the "bloat" you're referring to is because the Kindlegen output contains the MOBI version (for devices that don't support KF8), the KF8 version, and the source used to create the file (in this case, the content of your ePub).

EDIT: and before you ask, no... any per/MB delivery fees that Amazon charges you is not based on the size of the file that Kindlegen produces. Only one version (MOBI/KF8) is ultimately delivered to the user based on their devices/app's capabilities. The size of the files sent to the user is the basis for that fee. Mobi_Unpack (post #5) is capable of splitting the KF8 portion of the combined file from the MOBI portion if you'd like to get an idea of the file sizes that an end-user would actually be seeing.

[Jellby]

Originally Posted by Oldpilot

What does that mean? The ultimate "converter" is Kindlegen, surely?

What does "ultimate" mean here? All I dare say is that the final file created by Kindlegen should be a valid mobi (or KF8 or whatever). It doesn't mean it gives you the best possible representation of what the input file was supposed to look like. There is not a 1-to-1 conversion from ePub to mobi.

Note also that the "quotes" CSS property is not supported by ePub 2.01 (meaning that compliant reading systems are not required to support/recognize it), I don't know about KF8.

[_savage]

Thank you for the responses! After I posted the question and spent the evening reading random (seemingly) related threads, I found that CSS is not part of the .mobi file. Gah.

What seems to work: replace the <q class...> tags with <span> and style the spans appropriately; that then translates in text without the quotes. Both, kindlegen and Calibre seem to add the quotation marks automatically (independently of the styling) and I couldn't find how to disable that.

As for the statement above that kindlegen is the "ultimate" converter: well, it did a very poor job with my Pali characters, and I dropped it immediately. I tried Calibre and that hasn't messed up things. Yet :-)

What about the bullet lists?

Cheers!

[DiapDealer]

What about the bullet lists?

Lists have always been problematic in the mobi format.

Basically, if you have a "list" of things (like a ToC) and you don't want them to have any bullets (or numbers for that matter), then you probably shouldn't be using html list elements (<ul>, <ol>) in the first place. They don't really gain you anything (heck, I don't even use them when I do want something numbered or bulleted). Just use paragraphs. Style those paragraphs with margin-left (or use blockquotes) if you need indentation.

<ol>, <ul>, and <li> and all the css attributes that can be recognized/utilized (or will be recognized/utilized) varies so wildly and renders so differently on different devices/apps (ePub AND MOBI both) that I just don't waste my time with them.

[Doitsu]

Originally Posted by _savage

As for the statement above that kindlegen is the "ultimate" converter: well, it did a very poor job with my Pali characters, and I dropped it immediately.

The character issues might have been caused by incorrectly encoded .html source files. E..g non-Unicode utf8 files.
AFAIK, Pali can be written in a number of different alphabets. What kind of alphabet did you use and which characters did KindleGen have problems with?

[_savage]

Originally Posted by Doitsu

The character issues might have been caused by incorrectly encoded .html source files. E..g non-Unicode utf8 files. AFAIK, Pali can be written in a number of different alphabets. What kind of alphabet did you use and which characters did KindleGen have problems with?

Hmm.. Thanks for the hint! I just checked and noticed that none of my .html files had the appropriate header

<?xml version="1.0" encoding="UTF-8" ?>
<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.1//EN" "http://www.w3.org/TR/xhtml11/DTD/xhtml11.dtd">
<html xmlns="http://www.w3.org/1999/xhtml" xml:lang="en">

but instead only a stripped down version

<html xmlns="http://www.w3.org/1999/xhtml">

After I changed the headers, kindlegen produced a .mobi file with decent Pali encoding, and even the bullets of the lists are gone.



Now: Can I strip the .mobi of all the overhead that is embedded in it?

[Oldpilot]

Originally Posted by Jellby

What does "ultimate" mean here?

AFAIK, every file that is uploaded to the KDP is run through a converter, and that converter is Kindlegen.

[DiapDealer]

AFAIK, every file that is uploaded to the KDP is run through a converter, and that converter is Kindlegen.

Only files that aren't already in the MOBI format (including the combination MOBI/KF8 files) are run through a converter. If you upload html or ePub, then yes, it's going to get run through Kindlegen. But if you use Kindlegen (or MobiPocket Creator or calibre) to create your MOBI yourself... they don't convert that again.