Roam is a fantastic tool for thinking and writing in academia – it excels at helping you to synthesize your literature, thinking through problems and keep your writing on track.
Over the last year, I've taught hundreds of academics how to use it for their research, and the recently launched community for academics, @AcademiaRoamana, has now over 400 members. Interest in using Roam for science is strong, obviously.
But, of course there's a but!, there is one way in which @RoamResearch remains a HORRIBLE solution for most academics. I am talking about export. Let me explain why using "horrible" as a word is not hyperbole.
Leaving issues of collaboration and rounds of reviewer feedback aside (big thing, but bear with me), the academic writing process looks like this: you collect references using a reference manager like Zotero, you cite the works in your writing,...
you format your article to the specific requirements of the journal you want to submit it to, and then you submit your document as either a .docx file or (much less frequently) as a .tex or .pdf file.
I have to emphasize that using .docx files is _not_ a choice made by backwater scientists who haven't seen the light yet or are too old to change their ways. It's _required_ by the journals we submit to. No way around it.
Collectively, researchers around the world spend _thousands of hours EACH DAY_ formatting their .docx files for submission. Getting the citation format just right (every journals has their own preferences) likely has cost the world lifetimes of researcher output.
Those who are savvy have figured out that you can use tools like Zotero to insert references dynamically and have it format your bibliography later (still not enough people actually do this, wasting time doing it manually). Plugins for MS Word exist for this.
BUT these tools only work when you use them directly in Word – you can't open a Markdown file in Word (which you can't do anyways!) and then have Zotero find all the references – you need to start in Word. Markdown files don't open in Word at all.
That means that _even if_ you use the _amazing_ ZoteroRoam plugin from @AlixLahuec _and use citekeys_, simply exporting the page from Roam leaves you SOL. You'll have to manually go through and convert every reference (and footnote) after copy-pasting the text into Word.
You can, if you're savvy and/or so lazy you like to spend hours looking for solutions, figure out how to install Pandoc to convert your .md file into a .docx file. This is unlikely to happen unless you have someone like me showing you how to do it.
There is, of course, the LaTeX crowd – who voluntarily use LaTeX (most of the time) and therefore are at least familiar with CLI tools like Pandoc or have solutions like Overleaf that make that whole ordeal easier.
LaTeX users are the Linux users of desktop publishing – if you use LaTeX I can confidently assume you know how to deal with package managers. Do not assume all scientists are willing and/or able to convert to LaTeX.
The LaTeX users can deal with the "get markdown file run pandoc" workflow – but even for them this introduces hassles and hurdles that take time away from actual research and writing.
That's because we're substituting one hassle (format things for journal) with another: export Markdown file and run pandoc. Or (if using Pandoc is too hard) with "copy paste to Word and replace citekeys manually".
This suuuucks – and is the one major complaint @LaptopLifeLisa and I get when teaching academics to use Roam. Getting stuff out of it in a way and format that's useful is too hard. Did I mention to make the pandoc process work you need Roam42? Native Roam sucks for it.
So, how can @RoamResearch get a 10-100x ROI on dev time invested? Make actual academic writing and exporting easy. Do that, and you'll win over academia in a heartbeat. Multiplayer graphs for collaborative writing (if desired) and then export to Word and/or LaTeX with one click.
What should this look like? Some suggestions:
Let me use citekeys – either through Zotero API or .bib file upload. Let me define citation style through CSL files (Zotero API or repo). Let me use pandoc-style footnotes (works great with block references or (()) btw) and images.
Fully formatted output to Word, flattened and clean export to Markdown and .tex so I can pipe those through pandoc myself with my own templates or sync it to Overleaf.
I am, as always, ready to jump on a Zoom call and hash this out and demo in detail if you're ready to take this seriously, @RoamResearch. Say the word and I'll be there, day or night.
Cortex Futura -- Out.
• • •
Missing some Tweet in this thread? You can try to
force a refresh
So I'm working on a video on context switching, interstitial journaling and related things – things I learned from @ultraworking and how I use @RoamResearch for it now. I'll thread my brainstorming for the video below, feel free to ask questions. #roamcult
We all know that context switching is bad, Maker/Manager schedule etc. Fact is, we still have to context switch all the time, even if we have control over our schedule and work.
Projects take more than one day or block of hours – so you inevitably have to switch "in and out" of a given project. Even if that switch is just between personal life and that singular project.
If you've just recently discovered Roam, and you're checking out the community, things can feel overwhelming. For an app this young, the ecosystem is huge: YT tutorials, courses, extensions galore – plus regular new features in the app.
And on Twitter in particular, you'll often see people discuss the newest extensions, talk the "meta-game" of note-taking and Algorithms of Thought or celebrate the (fantastic!) submissions to the #RoamGames.
Even in the first 20% of the book it‘s _completely_ obvious that @zsviczian is right when he says De Bono has been working on Algorithms of Thought for decades, under a different name.
I feel like I‘m Neo following Morpheus into the Matrix
“All questions are attention-directing devices. We could easily drop out ‘questions’ and instead ask people to direct their attention to specified matters.”
– that‘s an important piece of figuring out how to think better, imo.
[[Algorithms of Thought]] are useful because they provide actionable guidance for how to deal with a specific problem without requiring a detailed, fixed checklist in advance. They can guide your thinking without being prescriptive.
The power of [[Algorithms of Thought]] can be even greater when you find ways to _externalise_ them, i.e. do not have to rely on memory to work through them. Memory alone can work well when the algorithm is fairly short and simple – but /
What are the [[Fundamental Skills of Knowledge Work]]? What do you think, #roamcult?
Boring but almost certainly the most undervalued skill by people typing on keyboards for a living the world over: typing speed.
Common objection to that: "My typing is irrelevant, it's the thinking that counts". Wrong. If a basketball player has to consciously think about how many steps they can take while dribbling, they get nowhere. If you need to look at your fingers to hit the right keys – fix that.
[[Knowledge Management]], [[Reproducible [[Social [[Science]]]]]], and [[Academic Workflow]]s – 100 Tweets for @threadapalooza 2020, let's go #roamcult#𐃏
1/100
Pandoc is a magical piece of software, and if you're not using it for your academic writing you're missing out. Compile (basically) any document format to (basically) any other document format.
2/100
While Pandoc is fantastic, it's a bit like ffmpeg: extremely powerful, but without GUI apps too few people will use it. ffmpeg has a ton of GUI apps that basically just wrap the CLI, Pandoc doesn't have enough of them.
3/100