Writing down my current process to better understand all the details of what I do when, why I do them, and what the current frustrations are as well as what can be improved with better tooling. Hopefully, this will generate plenty of ideas of what to build and prototype, and also give a sense for what’s most important and promising.
I wrote most of this in June 2019 and some minor parts have changed since then although the overall process is largely similar. You will see a fwe pointers marked with -> which are cross-links to other notes in my note taking system, which I have copied this from. I thought it’s a quick and easy task to publish this in its current form to get a general temperature read on how much our individual goals, techniques, and identified shortcomings overlap or differ from one another. If you have thoughts on this you’d like to share, please get in touch.
If I discover something I find worthy to spend more time with (reading, studying, watching, or listening to it), I put it on my reading list, which currently lives in Things. I do this to make sure I won't forget about it, and to eventually spend time with it. I have also discovered that putting it onto a list first instead of consuming the piece of content right away, lets me (theoretically) pick content to consume based on priorities after the excitement of discovery has cooled down a bit and I’m no longer biased on what occurs in the moment.
Each website, video, podcast, etc. becomes a task in Things. The title is (hopefully) enough to remind me what it was about, sometimes the link helps too, mostly because the source domain tells me more about what to expect.
I often miss recording more useful information and later wish I had added it:
- Where and why I found this (What was I doing? Was I searching the web? For what? Was I browsing Twitter? Was it coming up in conversation? Did somebody recommend it?
- The person who recommended this to me
- The time and date (and place?) I added this to my list
I sometimes add a little note on how many pages a paper has, or what the playtime of a video is, to get a feeling for the level of commitment I need to make when consuming that content.
For resources that require a significant effort to work through (e.g., books), I like to check ratings and reviews (for books on Goodreads, maybe Amazon and Apple Books) as well as availability and price (on Apple Books, my platform of choice for books).
I also like to investigate who published the content and when it was originally published, to fit it into a larger context of relevant people and historical context, or where in the author's timeline a piece belongs, if I know other pieces by the same person.
I realize that first of all, I’m using a task manager for something that is more of a resource collection, which isn’t ideal because it adds more stuff to my to do lists. It’s the wrong place and the wrong context, suggesting a wrong process driven by the wrong mindset. And second, even most of my common to do items are often not really clear-cut actions that need to be done and mustn’t be forgotten, but rather a collection of ideas, opportunities, and possibilities to choose from when I’m in the right mood and mindset.
An app that helps me “manage” both my reading lists and my to dos needs to provide all the context I need to make progress on a specific project. What I need more than a list of potential actions to execute on is a presentation of all the open questions and already found answers — with that I can pretty easily figure out myself what the next step should be.
I acknowledge the existence of straightforward action items that need to be recorded to not be forgotten and that are simple, clear-cut tasks that just have to be done. And the app should offer the appropriate functionality to manage these too. It does seem to be a smaller use case, at least for myself. Which means I could probably also just leave that to an app like Things or Reminders, instead of adding bloat.
–> Kinder To Do Lists
Something like Trello is a good solution for a slightly less task-oriented and slightly more resource-oriented workflow. A lot of small features from there are very useful. It also provides a slightly better overview thanks to its board view, however, that is also its weakness because it is the only way to present information in Trello.
As my reading list keeps growing faster than I can consume the content in it, I need a system to help me select the 20% that deliver 80% of the insights I want or need. This is why I need an overview of what’s on my list, without getting distracted with too much detail — sort of a map guiding me through my collected resources.
I have a few very broad categories of reading lists, as separate projects in Things, currently:
- General
- Productivity
- Complexity
- Language Design
- Programming
- Functional Programming
- Data & Persistence
- iOS Development
- Data Visualization
- Design, UI/UX, Art
- Graphics Programming
- Game Development
- Business
- Networking
Newly discovered content goes into Things' inbox, until I find the time to assign tags to it:
- What kind of material is it? To read (article, paper, book), to watch (video), to listen to (podcast episode); podcast episodes don’t make it into Things at all because I add them to Overcast directly.
- I try to add a flag if this is something I really want to come back to soon (either ‘next’ or ‘shortlist’), but using that tag hasn’t really worked out well so far.
Once tags are assigned, I move each item from the inbox to the appropriate reading list project. Then I jump into the project to further move the item under a specific headline into an appropriate section. Often I'm not sure if I'm putting it in the right spot as I'm missing an overview. Sometimes I have to create a new section because nothing else fits really, but I try to only do that if I already know I have more than one item that should be in that section. Of course, that process is cumbersome and broken for most of my reading lists, where I now have accumulated resources moved from the inbox, but not further sorted under a headline.
This is still too complex, but I've experienced the benefits of having a more structured reading list first hand when I moved from a simple Safari reading list, so I'm okay with putting in the extra effort for categorizing each item.
Categorizing and tagging do feel like effort and because the set of categories is open and unbounded, defining categories is difficult and often turns out to change later. This structure is useful for a better overview, but with each new category and each new item assigned to one, the structure becomes harder to change. Unfortunately, the structures in my mind evolve quite quickly and so I end up with an old structure in my tool and a no longer matching structure in my head. It would be huge, if the structure of categories in my tool was much more lightweight and almost effortless to adapt to the changing structure in my head.
When I havemake time to read, I need a way to make a good decision on what to read next. Several conditions affect, what "good" means in that specific context:
- How much time do I have? Do I just want to quickly read an article on the go (<5 min), do I want to spend a proper block (<20 min or <40 min plus note taking) on a longer article or paper, or am I trying to make daily progress on a book or paper?
- What's my environment like? Am I on the train, reading on my phone? Am I on the couch with my iPad? Am I at my desk? Am I doing some other mindless task like laundry (perfect for a podcast) or do I want to watch a video while I'm eating? This ties into the expectations I have regarding note taking — do I expect to take notes?
- What category and sub-topic do I want or need to learn more about? This is basically, how my current reading lists are structured: one list per category, each has sections (headings) for sub-topics with tasks for individual materials. I tend to pick a few topics I work on in parallel, so items from those sections are more likely to be picked than others.
- What mood am I in? What am I excited about currently? Sometimes I just thought about a specific topic and am all excited about it, so it might just be best to read something about that particular topic instead of trying to focus on something else.
- What is likely going to be the highest-quality material to move me forward in one of my projects? Even though my reading lists are already filtered, the content quality still varies significantly. It's obviously hard to assess content quality beforehand, but it is often possible to order items by e.g., expected level (introductory, tutorial, comprehensive overview, specific how-to vs. broader principles, detailed analysis, etc.) and depth, for instance its probably good to start with a short introductory blog article before diving into a much more detailed thesis, if I'm not yet sure if the topic is interesting or useful.
Ideally, the structure of my list (or an overview of all my material) allows me to minimize the time spent browsing through what's on offer and jumping directly into the material.
Sometimes I can qualify a piece of material as either "seems interesting, if I ever get to it" or "this I really need/want to look at soon", so there is also some sense of priority for some content, but it's hard to define and even harder to make work within my current system.
It would be fantastic if, at least for research and learning specific things, the system could suggest the most appropriate next document. A sense of making progress would also be great. Currently progress is hard to visualize because I generally add more items to the lists than I check off.
My current system does also not explicitly support the transition from past to future, i.e. material to be read to material that has been read — Things keeps a log of completed tasks, which is basically useless though (which I don’t consider Things’ fault, as I’m using a tool for something is hasn’t been designed for).
I try to create a separate note for each piece of material that resonates and sparks ideas, but I currently neither keep any other record of the source material than a URL and copied paragraphs for parts I deem important enough.
This is where I move from the higher-level "big picture" view down into specifics, where I allow myself to focus on one particular piece of content with the primary goal of understanding the author’s messages and intent.
It all starts with a gorgeous, distraction free reading experience. If the source of the material is a web page, I’d like it to be reformatted into a nicely type-set reading view that removes all navigation and chrome and distractions (ads), and can be trusted to keep all the relevant content (lists, tables, images, etc.). Ideally, it should also be available offline. These days I’m mostly comfortable with Safari Reading View, which doesn’t quite check all the boxes but is good enough for me. The current COVID situation also means that I don’t care about offline availability at this time.
To help understanding the material, several techniques beyond just linear reading have proven useful:
- Understand the structure first: skim the table of contents first, see what the agenda of a presentation is, read the summary at the very beginning or the conclusion at the very end (or both) first to understand what the author's objectives are.
- Highlight sections and phrases that resonate for any reason. I switch between highlighting excessively and not highlighting at all a lot; still haven’t figured out how useful it actually is.
- If phrases or paragraphs spark ideas, I copy them into my notes or annotate these sections, if I’m reading a book in Apple Books. I usually copy or even transcribe content into my notes theses days to have everything in one place.
- Make structure visible: mark topic transitions, less important sections ("just an example for the concept explained earlier"), lists and enumerations of concepts, hierarchy in arguments and concepts.
- –> Robert Horn's 40 structured writing information blocks of seven different types
- –> Tagging
- Rephrase and summarize important points and insights in my own words, as if I am trying to explain them to somebody else (–> Feynman technique).
- If I wanted to prepare content for spaced repetition, creating that content (individual cards for review with a single fact to remember) here would be ideal.
Can I measure and visualize progress? How much of the source material went through these steps? Maybe it's enough to show how much has been highlighted, annotated, and what the original word count to summary work count is?
There is a subtle phase transition from understanding a particular piece of material and integrating its insights into my existing knowledge landscape to making broader connections between thoughts, also beyond this particular material. This can also be felt strongly when the current note representing this particular piece of content is no longer the best place to store a broader insight and deserves its own note.
- –> Note taking = several use cases in one
- –> Refining Process in Thoughts
- –> Learning as transforming hierarchies
- –> Better notes by answering questions
True learning comes from tightly integrating new material into an existing network of knowledge. Once the source material has been added to the tool, is understood, and derived insights have also been captured, it needs to be connected to other previously existing nodes in the knowledge network, to increase recall and enable discovery of further connections.
–> Adding value by adding structure
Reviewing is a step to expose myself to notes again that I took earlier but might have forgotten. If I felt they were important enough to capture in the first place, there should be benefits in resurfacing some of them every once in a while. This has been proven extremely valuable on several occasions when I was reviewing old notes so far.
When reviewing existing notes, several small actions are helpful:
- Just reading them might spark new ideas or new connections not made before.
- Little mistakes (spelling, grammar, choice of words, structure, etc.) can be corrected, some sentences rephrased, leaving the note in slightly better shape than before.
- Adding new connections to material that I added later, but I now recall and which seems related.
- Collecting notes in newly formed categories, or summaries, to add more paths to find and discover them.
- Splitting a large note into smaller ones. “Refactoring” to make individual insights stand out clearly.
- Adding different visualizations, examples, references, or other supportive material to make this note easier to use in a model/story/artifact (see below).
Although resurfacing random notes for review is a valuable option, ideally there are some heuristics in play that try to take current context into account and try to surface those that might fit even better to what I am currently thinking about.
It would also be great to integrate reviewing notes as an important part of my daily routine, partly replacing consumption of new material.
- –> How thoughts can evolve
- –> Augmenting long-term memory
- –> Note taking use cases, derived from my Markdown/CLI prototyping
- –> Accumulating positive change
- –> The Design Squiggle
Curation is becoming a big part of creating valuable forms of output. Even without synthesizing new insights from existing ideas, just collecting, pruning, and sharing carefully selected content sources and/or providing an introduction to a topic can be extremely valuable and worth spending time to put together.
When I think about creating some artifact, I often look to build a sound model first. Once I have a good mental model, producing a story from it, e.g., a presentation or an article, is easy. Coming up with an insightful model is the difficult part.
What I call model here, is a mental system of individual parts that form a coherent whole which creates insight. Understanding the model means narratives can easily be derived from it. Understanding comes from understanding how the components interact with each other.
When building (searching for) a new model, I often start from potential components that might or might not become part of the model, and I investigate the connections between them.
This is all described in a very abstract way, which makes it generic, but way too hard to understand. Need specific examples!