¶ Summary
- This article describes my current workflow for doing incremental writing.
- It also touches upon incremental reading because to me the two processes exist as part of the same cycle of destruction and creation.
- I use SuperMemo for incremental reading and spaced repetition, Obsidian plus my incremental writing plugin to do my writing and I use Wiki.js to publish my articles.
¶ See Also:
- My summary article of incremental writing.
- A video where I commentate a session of incremental reading and writing.
- Piotr Wozniak's article on incremental writing.
- Destruction and Creation: My summary of a paper by Colonel Boyd that inspired my incremental writing workflow.
¶ The Cycle of Destruction and Creation
Incremental reading and incremental writing are opposite processes in the sense that they proceed in opposite "directions": in incremental reading you break down articles into extracts and flashcards, while in incremental writing you take extracts and flashcards and build them back into articles! But they are also complementary in the sense that incremental reading creates the extracts and flashcards that you will later incrementally recombine into new articles in incremental writing!
Rather than treating incremental reading and incremental writing as separate, I view them as two process entwined in an ongoing cycle of destruction and creation.
In this article, I will quickly sum up my workflow for incremental writing, by walking you through what I do for each stage in the cycle of destruction and creation.
I divide the cycle into 4 stages:
- Destruction
- Creation
- Publishing
- Recycling
- Taking finished articles and the feedback received from others and adding them back into my incremental reading queue.
¶ Stage 1: Destruction
The first stage in the cycle is one of destruction: using incremental reading in SuperMemo, I import input material in the form of books, articles and blog posts into a queue where each queue element is prioritized according to its expected value.
When I do incremental reading, I am like a knowledge goblin with a pickaxe, chipping away at an article or book, looking for the golden nuggets that I can extract and refine into atomic flashcards to be reviewed over time with spaced repetition. I think of the destruction phase as the "ascending" phase of the process: the number of individual pieces of information you have grows as the input articles and books are decomposed and split into fragments.
SuperMemo has a bunch of functions that support this process:
¶ Spaced Repetition
- SM18 supports inter-interval reviews well.
¶ Prioritisation
- Priority queue
- Batch prioritisation
¶ Subset Reviews ^eflz6zs
- Create tasks to remind yourself to review
- Search and review.
- Concept review.
- Branch review.
¶ Neural review
- Concept linking
- Model building and breaking.
- Input Output
¶ Stage 2: Create
The second stage of creation proceeds in the opposite direction to the destruction in the first stage: using my incremental writing plugin for Obsidian, I re-consolidate disconnected fragments into sentences, sentences into paragraphs and paragraphs into articles.
Incremental writing can be thought of as the opposite of incremental reading. The knowledge goblin takes the golden nuggets it mined while incremental reading and throws them into the forge to be melted down and mixed together with other ideas and concepts. Gradually, individual nuggets blend back into a comprehesive whole, and at the end of the process you are left with a solid gold bar of an article that you can share into the world.
¶ Importing from SuperMemo
- I am considering writing a plugin to do this.
¶ Linking
- Adding links between concepts that help readers gain the necessary prerequisite knowledge to understand your ideas.
¶ Reviewing Incrementally
- Using my incremental writing plugin I can review articles and individual "blocks" within articles over time using simple A-Factor scheduling inspired by the topic spacing algorithm in SuperMemo.
"A-Factor is interpreted as the number by which the current interval should be multiplied to get the value of the next interval" - Piotr Wozniak.
¶ Prioritizing
Just like in SuperMemo, each element in the queue is prioritized according to its importance. This way, even if I don't finish my entire incremental writing queue every day, I still know I allocated my limited time and attention to the highest value writing tasks.
¶ Stage 3: Publish
¶ Stage 4: Recycle
- After I have published my articles into the real world and received feedback, I can reimport the article back into my SuperMemo collection.
- As of March 2021, I haven't got too much experience with this yet. Check back later for more thoughts! (I added this to my incremental writing queue
)