Scientific Blogging with Pelican and Pandoc

First of all a disclaimer: I am no expert in scientific blogging or even blogging in general. I have only participated in one other blog which failed miserably due to a lack of content.

Formal Requirements for Scientific Texts

The format of scientific texts is highly dependant of the respective field as well as the publishing media. However, there are some universal standards such as the IMRAD (Introduction, Methods, Results and Discussion) format for reports of original data [1]. Prerequisite for scientific texts can be gathered by taking a look at traditional papers and abstracts:

1. Structure of Text
   It should be possible to structure text logically. This includes different types of headings, paragraphs as well as the ability to alter font characteristics to some degree.

2. Support for Citations
   One of the most important parts of working scientifically is to not just state an opinion, but to back it up with the preliminary work of other parties.

3. Support for Figures, Tables and Equations
   There should be a simple way to create figures, tables and equations as well as their corresponding captions.

4. Support for References
   Automatic cross referencing as supported LaTex ("\ref") and most word processors should be available.

Currently Pelican (v.3.2) is able to satisfy the requirement for structured text as well as proper figures, tables and equations. However, it lacks support for citations, which is comprehensible, as this feature is usually not required for blogging. Nonetheless, it is required for scientific texts.

Pelican and Pandoc

As Martin Fenner stated in his call for scholarly markdown [2] most of these features are already supported by Pandoc. This covers an excellent implementation for citations, including support for the Citation Style Language (CSL) which is especially useful when writing papers or abstracts. The natural fit is to use Pandoc as backend for Pelican. This can be achieved by creating a Pelican reader plugin, which is fairly straight forward. An implementation of such a reader plugin can be done in well under 50 lines of code: Pandoc_Reader.

Lookout

Although Pandoc has excellent support for structured text, figures, tables, equations and citations there is room for improvement. This includes fairly simple issues such as cross referencing support using html anchors as well as more complex ones such as support for internal links for tables and figures.