Collaborative scientific writing tools

While procrastinating writing my own paper, I started to look around for the best way to write an academic paper (priorities, I know…). Currently, my writing routine most often involves using vim (with the vim-pandoc plugins) to write markdown and convert to PDF with pandoc. Everyone else in my lab uses MS Word or Google Docs, so when we collaborate, it is usually all in Google Docs. My ideal setup would allow me to write in markdown and my collaborators to contribute comments and changes via a WYSIWYG interface. I would then be able to view their feedback through a file version history or some markup syntax (such as CriticMarkup. In a perfect world,

• Comments should be displayed in a side bar like Google Docs, without the need to click them individually.
• Editing the file should be possible via a link (preferably with a password or login) without having to share the file separately via Google Docs / Dropbox, and not require the installation of any software.
• The editor should preferably support the Pandoc markdown spec or similar and support html-comments in the markdown files that do not show up in the WYSIWYG view.
• Focus should be on making the barrier as low as possible for the people I work with to adopt this workflow. I do not mind jumping through a few hoops to set it up.
• The rich text version could be limited to receiving feedback through track changes and comments, which could be rendered in markdown as CriticMarkup, GitHub diffs or similar.

I have been trying a few tools for this, including

• Prose, edits markdown files in GitHub repositories. No rich edit, track changes or commenting.
• Overleaf, Latex and semi rich editing. Synced to a GitHub repository. File history instead of to track changes. Commenting in rich text shows up after the paragraph instead of in a side pane, and is not available in the Latex files.
• StackEdit, markdown semi rich editing. No track changes feature, comments need to be expanded one by one instead of always displayed in a side panel. Must share a file via Dropbox to collaborate instead of just a link a link.
• Sharelatex, all editing in Latex.

After almost giving up, I remembered a tool that I checked out a while ago, Authorea, but that was missing a few features at the time. It turns out to have matured a lot since last time I tried it:

• Syncs to one of your GitHub repositories.
• Allows writing to be done in Markdown, Latex or HTML. All articles are rendered in a WYSIWYG interface online.
• If you just want someone to leave comments on your article in the nicely formatted web view, you can create a public URL to share with them. For editing access, they are required to create and Authorea account (but not set up git or anything like that).
• Comments are shown in a side panel like Google Docs, and the commented section is cited in the comment.
• You can have 1-3 private articles and sync them with your private student repositories on GitHub (in the manuscript needs to be private).
• There is support for interactive HTML figures and Jupyter Notebooks! All of which can be contained in the same Git repo as your Authorea documents.
• They support citations via bibtex, inline searching of pubmed/crossref/DOI, and dragn drop from Mendeley/Zotero/Endnote.
• Supports html comments in the markdown file, although not inline.

There are two caveats holding me back from running around and converting everyone in lab to Authorea:

• There is no track changes feature that is intuitive for people used to MS Word/Google Docs. When I chatted with the staff, they said there are big changes to the track changes web interface inbound this summer, so I am looking forward to that!
• Once a new text cell/block has been created, the chosen text format cannot be changed. This is a pity, since it means that collaborators still need to be familiar with the format you have chosen. There are icons for “bold”, “cite”, “headings” and so on, but it is not as seamless as it could be. Commenting can still be done on the WYSIWYG rendered HTML page, it is just editing of the text that suffers. Again, their staff said that this was something they are working on, but not as immediate as the track changes features.

One minor annoyance is that comments are not synced to the GitHub repository and are only available online. I suggested that the comments could easily be synced via a separate text file, let’s see what they say.

If the changes that roll out this summer make track changes intuitive for people coming from a WYSIWYG workflow, that I believe that I will switch over to Authorea for all my collaborative scientific writing.