Re: Converting Docbook epub to Kindle shows warnings
Bob Stayton <[email protected]>
| Newsgroups | gmane.text.docbook.apps |
|---|---|
| Message-ID | <[email protected]> |
The index in my /Power Shift/ ebook does not use dl, dt, and dd in the
epub xhtml. I gave up on them and rewrote autoidx.xsl to output div
elements instead, with CSS to supply the indents. In a nutshell:
<dl> became <div class="dl">
<dt> became <div class="dt">
<dd> became <div class="dd">
That's it. Since the divs are already nested by the indexing templates,
it turns out that the only CSS I needed was:
div.dd {
margin-left: 2em;
}
I had intended to fold this back into the next release, so thanks for
reminding me about this.
By the way, if you liked /Power Shift/, you should check out my new book
/Solar Dividends/ which describes how to use solar energy to eliminate
poverty in the world.
Bob Stayton
[email protected]
On 3/19/2020 9:16 PM, Robert Nagle wrote:
> This is very interesting to me. I have tried for two projects to
> generate indexes that would work on Kindle. The indexterm content had
> both a primary and secondary level. Everything would look good in
> epubs on epub reading systems(and be valid, etc), but in mobipocket it
> would render inconsistently the elements in the index -- sometimes
> indenting, sometimes putting them on the same line. It was a mess.
> Even though it rendered fine in Kindle Previewer on all simulated
> devices, the dl, dt, dd just scrambled everything on Kindle app for
> android as well as on e-ink Paperwhites.
>
> I spent a lot of time trying to troubleshoot (and the details were
> fuzzy), but concluded that the problem lay not with docbook but with
> kindlegen (or maybe the css support on kindle reading systems). It's
> frustrating because having a good multilevel index is standard on many
> nonfiction ebooks.
>
> On the other hand, when looking at Bob's POWER SHIFT ebook (which came
> out a few years ago), the index rendered fine on Kindle. - although in
> that ebook, it used page numbers for an ebook.
>
> (In my projects, I used <xsl:param name="index.prefer.titleabbrev"
> select="1"></xsl:param> and <xsl:param name="index.links.to.section"
> select="0"></xsl:param> )
> (I could provide an ebook example if you need)
>
> BTW, never got around to saying it, but Bob I enjoyed Power Shift ebook
>
> Robert Nagle
>
>
>
>
>
> From: Richard Hamilton <[email protected]
> <mailto:[email protected]>>
> To: Lars Vogel <[email protected] <mailto:[email protected]>>
> Cc: Jason Zech <[email protected] <mailto:[email protected]>>,
> DocBook Apps <[email protected]
> <mailto:[email protected]>>
> Bcc:
> Date: Thu, 19 Mar 2020 18:42:50 -0700
> Subject: Re: [docbook-apps] Converting Docbook epub to Kindle shows
> warnings
> Hi Lars,
>
> I know this is from a long time ago, but I just ran into the same
> problem with a file that nests variable lists using the
> list-presentation=“blocks” processing instruction.
>
> In that case, the XHTML5 (this is with 1.79.2, building an epub3) is
> valid, but kindlegen doesn’t like the nesting. It closes off the
> highest level <dl> and <dd> elements and flags some <dt> elements as
> being invalid.
>
> The result doesn’t lose any content, but it removes the nesting and
> flattens everything to one level.
>
> As with your case, the epub is valid and when displayed, the nesting
> is fine.
>
> At this point, I’m convinced that the problem is in kindlegen. I just
> used the Kindle previewer, which wasn’t available in 2012, and it
> created a .mobi file that preserves the nesting (at least on the
> devices I tried in the previewer).
>
> Anyway, this is probably way to late to help, but I figured I’d add it
> to the thread in case someone runs into it in the future, since it is
> still a problem using kindlegen.
>
> Best regards,
> Dick Hamilton
> -------
> XML Press
> XML for Technical Communicators
> http://xmlpress.net <http://xmlpress.net/>
> [email protected] <mailto:[email protected]>
>
> --
> Robert Nagle
> 5115 Sandyfields Ln Katy, TX 77494
> (Cell) 832-251-7522