Additional Markdown we need in Scholarly Texts

Following up from my post last week, below is a suggested list of features that should be supported in documents written in scholarly markdown. Please provide feedback via the comments, or by editing the Wiki version I have set up here. Listed are features that go beyond the standard markdown syntax.

The goals of scholarly markdown are

  1. to support writing of complete scholarly articles,
  2. don’t make the syntax more complicated than it is today, and
  3. don’t rely on HTML as the fallback mechanism.

In practice this means that scholarly markdown should support most, but not all scholarly texts – documents that are heavy in math formulas, have complicated tables, etc. may be better written with LaTeX or Microsoft Word. It also means that scholarly markdown will probably contain only limited semantic markup, as this is difficult to do with a lightweight markup language and much easier with XML or a binary file format.

Cover Page

Optional metadata about a document. Typically used for title, authors (including affiliation), and publication date, but should be flexible enough to handle any kind of metadata (keywords, copyright, etc.).


Scholarly markdown should support superscript and subscript text, and provide an easy way to enter greek letters.


Tables should work as anchors (i.e. you can link to them) and table captions should support styled text. Unless the table is very simple, tables are probably better written as CSV files with another tool, and then imported into the scholarly markdown document similar to figures.


Figures in scholarly works are separated from the text, and have a figure caption (which can contain styled text). Figures should work as anchors (i.e. you can link to them). Figures can be in different file formats, including TIFF and PDF, and those formats have to be converted into web-friendly formats when exporting to HTML (e.g. PNG and SVG).

Citations and Links

Scholarly articles typically don’t have inline links, but rather citations. The external links (both scholarly identifiers such as DOIs and regular web URLs) are collected in a bibliography at the end of the document, and the citations in the text link to this bibliography. This functionality is similar to footnotes.

Citations should include a citation key in the text (e.g. [@Smith2006] or [#Smith2006]) and a separate bibliography file in BibTeX (or RIS) format that contains references for all citations. Inserting citations and creating the bibliography can best be done with a reference manager.

Cross-links – i.e. links within a document – are important for scholarly texts. It should be possible to link to section headers (e.g. the beginning of the discussion section), figures and tables.


Complicated math is probably best done in a different authoring environment, but simple formulas – both inline and as block elements – should be supported by scholarly markdown.


Comments are important for multi-author documents and if reviewer feedback should be included. Comments should be linked to a particular part of a document to provide context, or attached at the end of a document for general comments. It would also be helpful to “comment out” parts of a document, e.g. to indicate parts that are incomplete and need more work. Revisions of a markdown document are best handled using a version control system such as git.


This entry was posted in Thoughts and tagged , . Bookmark the permalink.

4 Responses to Additional Markdown we need in Scholarly Texts

  1. Chris Rusbridge says:

    This looks a great start, Martin. Two minor quibbles:

    – Typography. I suggest superscripts and subscripts should be supported in some semantic way rather than typographically. Putting semantics in via typography is a serious downer for the important business of reading the resulting article computationally (ie text mining). Also, there should indeed be support for different character sets, but again I’m not exactly sure this is a typographic thing.

    – Version control shouldn’t be dismissed quite so easily as the domain of git etc. At least some minimal level of version control should be included, as human readable version information is important when developing a document. Perhaps in conjunction with the Cover Page?

  2. Martin Fenner says:

    Chris, thanks for your feedback. With superscript and subscript I was referring to things like substance names. I like the simple way that pandoc is doing, e.g. 2^x^ for superscript and H~2~O for subscript.

    Adding a version to the cover page information is a good idea. Pandoc uses ~~is deleted text.~~ for deleted text. A markdown tag for the HTML element (not part of standard markdown) might be appropriate for text added since the last version.

  3. Carl Leubsdorf says:

    Very interesting and thorough discussion of efforts to consolidate markdown dialects and extensions: That post has also spawned a W3C Markdown community group, which might be relevant.

  4. Mathias Walter says:

    Regarding to the citations: A Markdown attribute for the Citation Typing Ontology (CiTO) would be useful and would probably encourage authors to use it.
    It is also important that the rendering engine (the converter i. e. to PDF/docx) supports Markdown/LaTeX commands in the BibTeX file. For instance, species names have to be written in italics. Pandoc or better, bibutils currently does not support this (see!msg/pandoc-discuss/Sr644VQU-Jg/px2jjMaCI-MJ).