Improving Pandoc Support for ConTeXt
Dear list, Pandoc, if you have not heard of it before, is a rather brilliant program for converting text between different markup languages: it allows, in essence, for technology-independent writing. It has been mentioned a few times on this list, with Pablo Rodríguez recently developing a set of templates for typesetting its XHTML output (at https://github.com/ousia/from-pandoc-to-context). One of its aims is to ease the process of producing a high-quality PDF; it has the strongest support for LaTeX at the moment, but it seems to me that ConTeXt would ultimately provide a more reliable and lightweight solution. The basic support is already in place; it only needs to be more thoroughly updated to take advantage of MkIV. I am working on improving Pandoc's generic ConTeXt template, at https://github.com/jgm/pandoc-templates/blob/master/default.context; the initial proposal is at https://github.com/jgm/pandoc-templates/pull/138, which aims to show more of ConTeXt’s potential to the user through making variables available that configure its options. I am still new to ConTeXt (having used LaTeX for several years), and could not find a recent starter document. The example at http://wiki.contextgarden.net/Detailed_Example is from 2006, and clearly does not exhibit the system’s full potential. I would be grateful to know of any potential areas for improvement. If you feel even more inclined to develop the Pandoc support for ConTeXt, its writer can be found at https://github.com/jgm/pandoc/blob/master/src/Text/Pandoc/Writers/ConTeXt.hs. In my experience, ConTeXt is much easier to use than LaTeX, but it is more difficult to get started with initially due to the lack of resources for new users; it is my hope that full support for it in Pandoc will provide one way of delving into it more deeply. All best, Andrew Dunning PhD Candidate Centre for Medieval Studies University of Toronto http://andrewdunning.ca
On Mon, 19 Oct 2015, Andrew Dunning wrote:
Dear list,
Pandoc, if you have not heard of it before, is a rather brilliant program for converting text between different markup languages: it allows, in essence, for technology-independent writing.
I used pandoc to generate class notes for a course (I needed to generate both HTML output and PDF output). I used the system for three years (I no longer teach that course), and I found pandoc to be very limiting, I ended up using a lot of preprocessing macros (https://randomdeterminism.wordpress.com/2012/06/01/how-i-stopped-worring-and...)
It has been mentioned a few times on this list, with Pablo Rodríguez recently developing a set of templates for typesetting its XHTML output (at https://github.com/ousia/from-pandoc-to-context). One of its aims is to ease the process of producing a high-quality PDF; it has the strongest support for LaTeX at the moment, but it seems to me that ConTeXt would ultimately provide a more reliable and lightweight solution. The basic support is already in place; it only needs to be more thoroughly updated to take advantage of MkIV.
This translation is not written in a ConTeXtish way. A lot of \startmode[*en,*uk] \chapter{...} \stopmode \startmode[*es] \chapter {...} \stopmode etc could be written using \labeltext. For example: \chapter{\labeltext{notes}] \setuplabeltext[en][notes={Notes}] \setuplabeltext[es][notes={Notas}] etc. I think that it will be easier if the pandoc-to-context code converts XML environments to ConTeXt environments and provides a default implementation of those environments. For example, <strong>...</strong>, etc. can be translated to ConTeXt macros (such as \pandocstrong{...}) rather than direct formatting commands. Then \pandocstrong could be defined as \definehighlight[pandocstrong][style=bold] Simplicarly, h2, h3, can be translated to \startpandocsection etc., allowing the user to change the level of the sections. Have a look at m-markdown in the source for such an implementation. This makes it much easier for a user to change the style.
If you feel even more inclined to develop the Pandoc support for ConTeXt, its writer can be found at https://github.com/jgm/pandoc/blob/master/src/Text/Pandoc/Writers/ConTeXt.hs. In my experience, ConTeXt is much easier to use than LaTeX, but it is more difficult to get started with initially due to the lack of resources for new users; it is my hope that full support for it in Pandoc will provide one way of delving into it more deeply.
I think that parsing pandoc's XHTML output is a more robust solution than changing the ConTeXt writer.
I am working on improving Pandoc's generic ConTeXt template, at https://github.com/jgm/pandoc-templates/blob/master/default.context; the initial proposal is at https://github.com/jgm/pandoc-templates/pull/138, which aims to show more of ConTeXt’s potential to the user through making variables available that configure its options. I am still new to ConTeXt (having used LaTeX for several years), and could not find a recent starter document. The example at http://wiki.contextgarden.net/Detailed_Example is from 2006, and clearly does not exhibit the system’s full potential. I would be grateful to know of any potential areas for improvement.
I find that it is easier to write a new template according to my needs rather than make the default template flexible enough that it fits everyone's needs. One of the big advatanges of ConTeXt, as compared to LaTeX, is that it is relatively easy to write a new style starting from scratch. Aditya
Aditya Mahajan mailto:adityam@umich.edu 24. Oktober 2015 um 21:11 On Mon, 19 Oct 2015, Andrew Dunning wrote:
It has been mentioned a few times on this list, with Pablo Rodríguez recently developing a set of templates for typesetting its XHTML output (at https://github.com/ousia/from-pandoc-to-context). One of its aims is to ease the process of producing a high-quality PDF; it has the strongest support for LaTeX at the moment, but it seems to me that ConTeXt would ultimately provide a more reliable and lightweight solution. The basic support is already in place; it only needs to be more thoroughly updated to take advantage of MkIV.
This translation is not written in a ConTeXtish way. A lot of
\startmode[*en,*uk] \chapter{...} \stopmode \startmode[*es] \chapter {...} \stopmode
etc could be written using \labeltext. For example:
\chapter{\labeltext{notes}]
\setuplabeltext[en][notes={Notes}] \setuplabeltext[es][notes={Notas}] etc. Why don’t you use \headtext?
Wolfgang
On Mon, 26 Oct 2015, Wolfgang Schuster wrote:
Aditya Mahajan mailto:adityam@umich.edu 24. Oktober 2015 um 21:11 On Mon, 19 Oct 2015, Andrew Dunning wrote:
It has been mentioned a few times on this list, with Pablo Rodríguez recently developing a set of templates for typesetting its XHTML output (at https://github.com/ousia/from-pandoc-to-context). One of its aims is to ease the process of producing a high-quality PDF; it has the strongest support for LaTeX at the moment, but it seems to me that ConTeXt would ultimately provide a more reliable and lightweight solution. The basic support is already in place; it only needs to be more thoroughly updated to take advantage of MkIV.
This translation is not written in a ConTeXtish way. A lot of
\startmode[*en,*uk] \chapter{...} \stopmode \startmode[*es] \chapter {...} \stopmode
etc could be written using \labeltext. For example:
\chapter{\labeltext{notes}]
\setuplabeltext[en][notes={Notes}] \setuplabeltext[es][notes={Notas}] etc. Why don’t you use \headtext?
Because I did not know about it :-) Aditya
participants (3)
-
Aditya Mahajan
-
Andrew Dunning
-
Wolfgang Schuster