Well, that's a bit disappointing. Thanks for the full report. I've only
used Pandoc for simple conversions, so haven't looked deeply at the
configuration options that might help with the problems that you
identify below.

One last approach might be interesting to try: What about LyX ->
(X)html, then process through Pandoc. And instead of using the LyX
exporters, try tex4ht to make the html file. And process the html file
through tidy before using Pandoc.

I haven't tried this, so please don't waste your time on it if
inconvenient. Pandoc seems at its strongest when starting with a
Markdown file.

Cheers,
Alan


Steve Litt writes:

> On Sat, 01 Feb 2014 05:25:58 +1100
> Alan L Tyree <alanty...@gmail.com> wrote:
>
>> Sorry for the top posting, but this is short. My own view is that an
>> ePub exporter for LyX would make it a killer application. 
>> 
>> Thanks for your comments, Steve. Have you looked at Pandoc?
>> 
>> Cheers,
>> Alan
>
> Hi Alan,
>
> I hadn't known about Pandoc. Thanks for the heads-up. I looked up Pandoc
> here:
>
> http://johnmacfarlane.net/pandoc/
>
> For the following ascii art, please switch to monospace font...
>
>            .___.               .______.
> LyXformat->|LyX|->LaTeXformat->|Pandoc|->ePubFormat
>            `````               ````````
>
> The preceding looked simple enough, so I downloaded and installed
> Pandoc, and tried it on one of my simpler books. It indeed produced an
> ePub, and in certain respects a good one. But not nearly good enough to
> sell. No table of contents. No cover. No images: all images referred to
> only by their Alt text. All cross reference labels exposed as text with
> arbitrary subscript formatting in places. The good news is it managed
> to keep footnotes and the like, but that's not good enough. I spoze
> theoretically I could have used other options on my lyx -export
> command, or on my pandoc command, so these things wouldn't happen, or
> perhaps I could have made restrictions on the authoring of my book in
> LyX, but these things would need to be examined later.
>
> Another way to use Pandoc might be this:
>
>            .___.               .______.
> LyXformat->|LyX|->LaTeXformat->|Pandoc|->xhtml-.
>            `````               ````````        |
>  .---------------------------------------------'
>  |
>  |  ._____________.         ._____________.
>  `->|xhtml Tweaker|->xhtml->|My xhtml2epub|->ePub
>     ```````````````         ```````````````
>
> The preceding would depend on:
>
> A) Pandoc retaining enough info, including semantic styles, to make all
>    book elements
> B) Pandoc producing xhtml sane enough that the tweaker is something
>    that can actually be written.
>
> So I tried using Pandoc to convert my LaTeX book to (X)html. The result
> had no <head> (and this might be an advantage), it had all sorts of
> garbage characters (perhaps this could be fixed in the <head> I would
> insert), but, the kiss of death is this: It took all my semantic styles
> (environments and character styles), kinda-sorta converted them to
> presentation, and discarded the semantic styles, so that in my <head>
> I can't specify the link between semantic styles and presentation.
> Once those semantic styles are gone, no matter how clever a programmer
> I am, I don't have the necessary input info to govern the look of my
> eBook. This is a showstopper that cannot be recovered from. 
>
> So unless somebody knows of a way to prevent Pandoc from pulling an
> MSWord move and prematurely converting semantic to presentational, my
> opinion is that Pandoc is worthless for this task.
>
> Thanks,
>
> SteveT
>
> Steve Litt                *  http://www.troubleshooters.com/
> Troubleshooting Training  *  Human Performance


-- 
Alan L Tyree           http://www2.austlii.edu.au/~alan
Tel:  04 2748 6206     sip:172...@iptel.org

Reply via email to