[NTG-context] ConTeXt epub workflow

Aditya Mahajan adityam at umich.edu
Thu Jul 3 05:42:22 CEST 2014

On Wed, 2 Jul 2014, Idris Samawi Hamid ادريس   سماوي حامد wrote:

> Dear gang,
> I have an urgent project that needs three outputs: pdf, epub, and kindle. The 
> formatting needs are pretty basic. In your experience:
> What is the recommended, more efficient workflow for this sort of thing? 
> Should I start with markdown and then generate context and epub (then convert 
> epub=>kindle I presume)? Or does one context file with pdf and epub outputs 
> work well?

If you are short on time and your formatting needs are basic, I would 
suggest that you start with markdown as your input format and use pandoc 
to convert it to epub and context. There are various tools to do epub to 
kindle conversion.

Markdown in a very primitive input format. For anything non-trivial 
(multiple types of floats, multiple types of emphasis, etc.) you have to 
resort to some sort of pre-processing of input.

Pandoc is the only tool that does Markdown to ConTeXt conversion. There 
are more tools for Markdown to LaTeX conversion. The ConTeXt markup 
generated by pandoc is not ideal (\bf, \em, etc instead of semantic 
markup); it is possible to change the output, but you need to learn how to 
program in Haskell.

You can start with ConTeXt and generate epub. However, ConTeXt generates 
XML+CSS rather than "pure" XHTML. Most browsers can handle XML+CSS, but 
most (all?) epub readers cannot. So, the output is not usable without some 
post-processing (this might have changed, I haven't tested ConTeXt epub in 
a year or so).

In the long run, I think that the "easiest" way to generate multiple 
output formats for non-trivial input text is to use XML as your input 


More information about the ntg-context mailing list