Master Format for multi-format eBook Generation?

[tompe]

Originally Posted by frabjous

But what about TeX > other eBook formats, like Mobi, ePub and LRF, etc.? Anyone have any experience with this, or know of any good tools/packages, etc., for it? How well do these support the features a TeX user typically expects, like decent looking mathematical formulae, footnotes, etc.?

latex2html actually works pretty well. It makes images to represent the math.I have seen one example of this converted to MobiPocket and it worked so and so. The images was the problem. They were to small and not so legible. But maybe that can be solved by some fine tuning.

[frabjous]

Thanks for the info. Do you know whether it create these images for everything put in "math mode" (to use the LaTeX terminology); because that would definitely be a problem.

[tompe]

Originally Posted by frabjous

Thanks for the info. Do you know whether it create these images for everything put in "math mode" (to use the LaTeX terminology); because that would definitely be a problem.

I think it might do that but you have to test it to be sure. But this does not have to be a big problem. You probably have to avoid long math words th like $\mathit{flies}(\mathit{tweety})$ in non displayed math mode.

[frabjous]

Thanks. Perhaps I'll do some experimenting when I get a chance, and report back.

[cerement]

Looks like there might be support out there, just trying to find it and find who's working on it. Short blog entry by Bill McCoy about Hindawi (a science/tech/math publisher).

Originally Posted by Bill McCoy

They are beginning to adopt EPUB, using LaTex-based workflow to generate SVG from MathML. ... The equations are "live" including selectable text.

[cerement]

Hmm, another item:

Threepress Consulting (of Bookworm fame) has a set of epub-tools available on Google Code, the two main tools are docbook2epub and tei2epub. One quick blurb mentions: " These tools have been used by individuals in ePub-generating projects and have been incorporated into the Calibre ebook conversion suite", although a quick search on Calibre didn't turn up any mention of TEI or DocBook ...

[frabjous]

Originally Posted by cerement

Looks like there might be support out there, just trying to find it and find who's working on it. Short blog entry by Bill McCoy about Hindawi (a science/tech/math publisher).

This is very interesting. For non-professionals, it looks like what you might try is to create something in LaTeX, convert it to GNU Texmacs format, and then convert that into xhtml with embedded MathML, and then convert that to ePub... a long process, but possibly resulting in better output than latex2html.

[Kirtai]

One thing I've noticed is that ePub directly supports books in DTBook format in addition to XHTML. Can anyone tell me how good DTBook would be for this purpose?
It would be really nice to be able to use a Master Format which can also be viewed directly by ereaders and has a full build toolchain ready made
(I admit to being something of a newbie at XML formats)

[cerement]

Originally Posted by Kirtai

One thing I've noticed is that ePub directly supports books in DTBook format in addition to XHTML. Can anyone tell me how good DTBook would be for this purpose?

DTBook is more focused on correct semantic markup. Not as freeform as XHTML. Found a couple posts by Jon Noring here and here comparing DTBook and XHTML (as well as a couple other XML formats).

[cerement]

Originally Posted by frabjous

... create something in LaTeX, convert it to GNU Texmacs format, and then convert that into xhtml with embedded MathML, and then convert that to ePub ...

Poking around and found a reference to TeX4ht (TeX to (X)HTML, MathML, etc), don't know anything about the quality of output.

[cerement]

The more I poke around teh intertubes, the more confusing the choices become. The biggest problem seems to be what features the various reading programs support. In our rush to separate semantic content from layout, readers have ignored or thrown out all the previous effort that has gone into getting great layouts (even on multiple devices). With the switch to XML, our readers have automagically forgotten all the programming that went into reflowing text, pagination, device independent output, hyphenation, and typography available in the TeX libraries alone.

Many publishers have put their name behind ePub, very few have published anything to ePub. In the meantime, they are more likely to use a hand-rolled XML format internally (even feedbooks admits to his ).

TEI looks interesting in that it allows an author to use inline TeX to express math formulas. The rest of the formats tend towards MathML. And DITA hasn't settled on a math domain yet.

DITA has the advantage of allowing very non-linear documents. As Jon Noring points out, just try to express a website or Wikipedia in ePub. In the Palm days, the database program TomeRaider did a very good job of handling both dictionary content and linear narrative. DocBook, TEI, and ePub all tend towards a very linear approach.

[Kirtai]

Originally Posted by cerement

The more I poke around teh intertubes, the more confusing the choices become.

No kidding. It's very confusing.
It's a pity the tools aren't mature enough to just let people use appropriate markups (e.g. DTBook for fiction, DITA for articles/magazines/anthologies) and automatically convert them to display formats like ePub.

Not to mention, actually displaying those output formats properly instead of half-assed (I'm looking at you, Adobe Digital Editions)

And then there's OpenDocument format, which a lot of organisations such as governments are switching to, which helps muddy things nicely.

[frabjous]

Originally Posted by cerement

Poking around and found a reference to TeX4ht (TeX to (X)HTML, MathML, etc), don't know anything about the quality of output.

That one looks rather complicated even to set up. I've experimented a bit with this: tex2xml, an online converter, and have been pretty happy with the results I've been getting, but I plan to do some more experimenting.

EDIT: Actually, while the xml looks good, it didn't yield great results when I tried to convert the xml to ePub with calibre... sigh... I'll keep playing.

[JSWolf]

Unless the eBook in question has need beyond the bounds of HTML, then HTML is the optimum format for multiple format generation. Calibre can take that HTML and generate ePub, LRF, LIT, and Mobipocket. So why would you need this nonstandard DTBook in most cases?

[Kirtai]

Originally Posted by JSWolf

Unless the eBook in question has need beyond the bounds of HTML, then HTML is the optimum format for multiple format generation.

I just find it annoying that marking up stuff in HTML is done by convention, rather than by actually marking things as they are.

Originally Posted by JSWolf

Calibre can take that HTML and generate ePub, LRF, LIT, and Mobipocket. So why would you need this nonstandard DTBook in most cases?

The ePub spec actually includes DTBook as an officially supported format, mainly for combined text/speaking books.