title | author | affiliation | keywords | numbering | |
---|---|---|---|---|---|
SciDown Manifesto |
Martino Ferrari |
University of Geneva |
SciDown, MarkDown, LaTeX, Publiscing |
false |
Honestly LaTeX is great, the quality of the output produced is outstanding, however its syntax is both dated and lengthy.
Moreover due to the compilation time of the documents there are no (at my knowledge) editor with (real) real-time rendering capability, making the creation of the document more unpleasant that it could (and should) be.
Finally today publishing for the Web is becoming more and more important and be able to produce documents as both as pdf and html is a big plus.
MarkDown is now becoming more and more popular, as LaTeX is a plain text language where the users do not have to think about formatting or styling. Moreover the language is very simple, intuitive and short (e.g. # section tilte
against \section{section title}
). However many features are missing to make such language ready for writing scientific publications, for example floating elements such as figures, tables and listings with or without captions.
The goal of SciDown is to complete MarkDown with some of this missing features, to be easily integrable in editors and views and compile in both HTML and LaTeX.
As today SciDown (and its C implementation) is already very capable, it generates both HTML and LaTeX (needs improvements) and adds many features to MarkDown:
- Header (
YAML
): title, authors, keywords and more - Table of Contents
- Abstract section (with optional keywords)
- Floating elements (figures, tables, listings, equations) with labels and captions
- Internal label references
- Including external files, both as source or as references only
- 2D charts as SVG and PGFPLOTS generation using charter
- Easy
CSS
styling
For the future is planned to improve both stability of the parser and output quality for HTML and particularly for LaTeX.
Improve the plotting capabilities and customization, configure the image size and alignments (e.g. sub-plots and multiple images per figure).
Finally would be great to create a new simpler styling language to be used for both HTML and LaTeX output, or (maybe even better) to compile the CSS
in a LaTeX
understandable format.