Reading Inbox

This an Inbox zero-like approach to managing sources. The goal here is prevent an “anxiety-inducing”, never ending build up of sources to read and process but just get siloed away. So, once an item is captured, it will either

1. get trashed (doesn’t look like it’s worth a detailed read after all)
2. get read in a serious fashion (i.e. write about what you read)
3. get read shallowly and filed in the reference library
4. (maybe) gets added to some other list like “recipes to be cooked”

This makes sense, and while certain things should be deleted at times, I like that there is still room for “filing” sources away for later. Step 2 and 3 are mostly what the system’s feed component has been for, storing relevant metadata on a particular source along with no notes all the way up to full processing. Step 1 is somewhat outside the scope of the feed, having more to do with how I decide what sources to add, and there isn’t an exact place for Step 4.

As Andy mentions, the goal with this inbox is to close open loops and ensure thoughts, ideas, and in this case, sources aren’t begging to be rehearsed within our minds, taking up valuable headspace. In many ways, these ideas should also be applied to TODO lists.

Relevant notes

Transient Notes

A daily note titled with the date is created every day. Andy uses this note to capture reflections, scratch work, and any other “dumping ground” material. Like the way I’ve been using my daily notes, and as we’ve seen many others using Roam research’s daily notes (e.g. Building a Second Brain in Roam, Roam Why I Love It and How I Use It), having this note type ensures there’s “always a place to put that thing.” This matches my Log and Thoughts pages copied to every daily note to ensure I can take down random things I think of, reflect on, or work on throughout the day.

These notes are ephemeral by design, but extracted to the writing inbox when there appears there’s enough substance to the scratch notes. He notes these notes can also be extracted as-is to evergreen notes.

Overall, the important job of the daily notes are to

1. Provide a frictionless note taking location for anything that may come up.
2. Encourage movement of important material to other note types, those notes worth “outliving the day”

Writing Inbox

Transient (daily) notes should be easy to capture, and later easily turned into permanent (evergreen) notes. This requires

1. A quick way to capture transient notes (i.e. creating that daily page and letting yourself dump anything and every to it)
2. A place to put notes for further development and ensure this queue gets reliably drained (back to the concept of closing open loops)

Here Andy makes use of a “writing inbox”, a queue for undeveloped ideas, snippets from daily notes, reading notes, etc. I suppose this could loosely be likened to my so-called “Memory Stack”, where I dumped little ideas and notes hoping they would later be addressed (they weren’t). Andy notes that he spends the first few hours of his day going through the writing inbox and hashing out new evergreen notes where appropriate, and deleting or archiving the others.

This seems to be an effective method for ensuring loose ends are tied up on informal notes and ideas. However, I’m curious what exactly Andy does to archive a note when it doesn’t find its way to becoming an evergreen note.

Evergreen Notes

As we’ve seen, most people only take transient notes. These are convenient, easy to write notes that accumulate without association over time. With tools like Roam research and the introduction of backlinks, it’s much easier to focus on building a network of notes that grows knowledge exponentially over time, constantly building on past work instead of becoming siloed away on pages. I’ve noted many times that this is how I felt working with Notion and most note-taking systems throughout my life, but with a bit of enlightening from Roam and my work on Vim-roam, I’m trying to pivot away from siloed notes and embraced a more informal, networked way of note-taking.

Simply put, evergreen notes are like zettels from a Zettelkasten. Evergreen notes should be

1. Atomic (keep it on topic but complete)
2. Concept-oriented (rather than by project, book, topic, etc)
3. Densely linked (use those wikilinks)
4. Prefer associative ontologies to hierarchical taxonomies

Differences between this system and a Zettelkasten

The main aspects of the evergreen notes are virtually identical to the desired properties of zettels in a zettelkasten. Some primary differences in the rest of the practice are:

• Methods for capturing scraps and sources, including the reading and writing inboxes. This also includes and spaced repetition used to aid work here.
• A different note taxonomy, including the progression of writing very early from daily notes, movement into the writing inbox, and then evergreen notes. Additionally, descriptive note titles are relied upon as identifiers instead of long numerical sequences.
• Contextual backlinking
• Integrating spaced repetition into notes
• Making the notes publicly accessible and integrating them into public conversations

Write about what you read

Andy suggests writing Evergreen notes as you read. There are a few nuances here, however. Firstly, it can be distracting to switch back and forth between reading and taking notes. Instead, make markings of interesting passages or write simple thoughts on paper or a temporary document while reading. These can be pushed very quickly to the writing inbox, where it’s okay that we have some temporarily incomplete notes. These notes/remarks, however, should be enough to give you the proper context when reviewing and remind you why that thing was particularly salient.

How this collection of reading notes gets refined can vary. Andy suggests taking a big picture view of the ideas presented in the reading notes, and then clustering them into logical piles. A loose process from here:

1. Write a broad note capturing the big idea of one of the clusters
2. Break this note up into fine grained notes, focusing more on the individual atomic ideas/concepts within the cluster
3. Connect these new atomic ideas into the system. Search for relevant past notes. Link with old notes and merge new content in as necessary.
4. Iterate on the remaining notes in the cluster using the information gained integrating the last note into the system. Use what you learned from that interaction to update other temp notes or remove those that aren’t necessary.

Relevant notes

Note Taxonomy

• Daily scratch/log notes
• Prompts and incomplete notes in the writing inbox
• Evergreen notes
  • Stubs for backlinks
  • Simple definition notes
  • Precise, narrow notes; can be framed as questions
  • Higher level notes i.e. notes that abstract over groups of others or personally defined terms
• Outline notes: pull together many related evergreen notes under one roof and more or less represent an outline of a topic

Note types outside the taxonomy: - Literature notes: titled after a source, meant mainly as a link to other durable notes and a backlink target. Andy calls these “outline notes but for someone else’s ideas” * I’ve seen both loose outlines of other sources, as well as more in-depth and descriptive notes about sources in Andy’s system. These notes don’t have to be super light. - Log notes: observations about a specific system or project over time. I tend to take this in my daily notes, and can view them as backlinks.

Executable strategy for writing

Here Andy describes a process for maintaining and ultimately making use of Evergreen notes that have accumulated over time.

1. Write durable (evergreen) notes while reading and thinking. This is both for thinking about random things (and noting them in daily notes as scratch most of the time) and while taking literature notes.
2. When a note is added, add a link to that note in an appropriate outline (creating one if there is not a suitable one already)
3. When it comes time to actual write a piece, or you want to develop a topic further, find gaps in that outline and create new (evergreen) notes to fill in those gaps.
4. Aggregate all notes from the outline into an initial draft, and iterate. As Andy mentions, with managed notes the task becomes more like editing than composition.

What to do with reference notes

There’s more on an individual page of his Literature notes are separate. Evergreen notes are responsible for being concept-oriented and a place where we do our own thinking, our own interpretation of material, building other’s ideas into our system of knowledge. They have few direct quotes or source material from others. However, source pages are mostly thoughts directly from the author in the material being digested, and while this is important for getting the proper context and understanding the author’s point, it should be kept separate our own “zettel” system. Andy notes it’s “hard to hear yourself think”, so we should be sure to clearly differentiate between a space for our own thoughts and those directly representing others.

Luckily this works quite well with my current system setup; I have zettels as a separate folder of notes from the direct feed notes or wiki notes. All subsystems can be referenced freely between each other, but the original idea was to keep these separate precisely for the reason Andy mentioned: so I could make it easy to distinguish between content directly derived from source material and that of my own thinking. This is good reinforcement for my design.

What to do with wiki notes

Does something like Category theory need to be broken up across zettels (i.e. would that be effective)? Or should it remain as a more “wiki” page, with a complete reference and definitions?

Questions and remarks

• What software does Andy use to write these notes?
  • Looks possibly like Bear or Roam

Spaced repetition

Andy notes that most spaced repetition systems discourage the development of evergreen notes. On the other hand, an evergreen approach might help you build ideas over time, but it won’t help you reengage with those notes and retain their content. So, to get the benefits of both systems, Andy employs the mnemonic medium, which allows a user to define spaced repetition prompts within their personal notes.

{Cloze deletion} could be used to {define terms or phrases to be hidden as parts of a notecard}. Additionally, traditional Q&A notecards could be created with

Q. This is a question A. This is an answer

The idea behind the system is that, in a manner not unlike Taskwiki, a scan is performed on a collection of notes to update or add cards to the spaced repetition system.

In some ways, maintaining evergreen notes (i.e. a Zettelkasten) approximates spaced repetition as, in theory, you frequently visit those notes of interest (through backlinks or regular wikilinks) and read or update them. This forces you to engage with these notes in a manner similar to that of a regular spaced repetition, but, as Andy mentions, it’s not a very efficient system.

Andy’s implementation of a mnemonic medium is described in more detail here. He notes the similarity between his own approach and that of Obsidian to Anki, a plugin for Obsidian that exports a custom syntax to Anki cards. This for me is a big part of the Vim-roam and Personal management system that I’d like to integrate. Note: can actually just use that obsidian_to_anki script since, after all, Obsidian just sites on my Markdown files anyway!

Main Takeaways

Overall, Andy’s system goes something as follows:

1. Push possible reading material to a reading inbox. Reliably drain this queue, deciding to either toss the source, skim it, read it full (and take notes), or place it elsewhere.
2. Have a place to write daily notes. This is a local dumping ground for the 24 hour period; any logging, scratch work, fleeting ideas, etc are to be put here.
3. As things come up, whether they are fleeting notes, reading notes, or whatever, push these to a writing inbox. This queue is reliably drained, turning partially written material into evergreen notes or deleting the note.
4. During the process of creating durable notes, add them to outlines so they can be easily used later for writing (while providing nice, complete, global overviews of topics where the notes come together to form something more).

Additionally,

• Keep literature notes separate. If taking notes while reading material, keep them somewhat brief, lightweight observations. You can keep a summary of notes or whatever else in that document for reference, but the main goal is to let the bulk of the value be soaked up in evergreen notes and integrated into our knowledge web as a series of atomic ideas.
• Write about what you read. Take light notes on paper or make notes of stand-out paragraphs while reading. In theory, these could also be more full processing pages like this one. Then you iteratively turn these notes and ideas into evergreen notes by finding ways to interpret the material and link it into your current web of concepts and ideas.
• Zettels can be titled with a question. Want the note be have an inquisitive topic, maybe one you don’t have the answer to? Put the question in the title. Can be an argument of sorts, where you argue multiple sides of the point.

Created zettels (as something of a test):

• Note-taking practices are generally ineffective
• Note-taking provides weak feedback