Anyone using Markdown for scientific article writing? Scrivener?

I have LaTeX fatigue. It looks like someone ate an article and threw up on my screen.
I would like to use Markdown, and I would assume pandoc along with it.
I need:

  • Figures with captions (and previews in the editor, I suppose)
  • References to those figures in text (See Fig. 3)
  • Equations would be nice, but I could embed an image
  • Citations, of course

I don’t think there are any Markdown editors that support all these features, so I assume I would use something like Typora or iA Writer, write in Pandoc-flavored Markdown, then generate a PDF when I want to see a preview (or used Marked2, I suppose).

E.g. something like Wouter Groeneveld’s workflow

It’s been a while since I used Scrivener, is the workflow the same? I was always a little bewildered by whether I should be writing Markdown, or using the built-in styling tools.

Thanks for any suggestions!

Lingua vomite

Generally speaking, Scrivener’s core experience remains the same, but this post might be of interest to you


Stay away from the light. Stay away from the light …



In all seriousness, as one deeply immersed in LaTeX for just about anything except inter-office memos, I am always cautious when someone tries to convince me that I can forego the “ugliness” of LaTeX syntax source files by adding an “easy” intermediate conversion step to some other markup language.

IIRC, Scrivener can export as LaTeX.

I will also be especially interested to hear the end result when you are still using the long, underscore-rich, naming conventions for file names and labels. This is a battle into itself.



Another part of this is easing the conversion to Word for my advisor’s feedback. I’m paying $20/mo for Acrobat, and it’s doing a \frac{3}{4}-ass job of it.

I like snake_case. It shows me the progression of my files through my processing pipeline. When I needed to reprocess beginning in the middle of the process, I knew exactly which files. The underscores are also easier to parse than CamelCase would be.

Right now, I am using Scrivener and compiling the text to a Markdown file. With it, I then generate docx, html, and LaTeX versions via Pandoc. All of this is actually automated, so I get these formats (+ PDF, if LuaLaTeX doesn’t throw any error back at Scrivener) right away. There is also some post-processing going on with the MD file before it converts to LaTeX and with the LaTeX file before it converts to PDF.

There are actually three ways to achieve these conversions: i. using a script in Scrivener’s post-processing pane, ii. using Pandoc defaults, and iii. using Scrivomatic (which I highly recommend).

1 Like

I’m using Scrivener with scrivomatic and working on a journal article.
I like the product and appreciate everyone’s contributions, but the settings are all over the place, both within Scrivener’s compile settings, the pandoc metadata, and pandocomatic.yaml. I’m probably forgetting others.
I really don’t see how non-technical people handle it.


I definitively see what you mean. I think people can accept that setting this thing up will take effort or just let go of it and head towards a simpler route (like docx export).

I try to think about Scrivener as just a tool to produce a Markdown file. One with a lot of room for complexity, of course, with section types, section layouts, styles, and so on. But the end product is transparent enough for me.

As a side note: you might want to check the “Compile for TXT” option. It allows for the customization of some things that the MMD export does not. To mention just two examples that I find relevant: footnotes (you can have them inline using MMD or Pandoc syntax) and links. It will work precisely like the MMD export (even using the md extension) but it has to be in the txt pane, for some reason.

1 Like