What are good strategies for reviewing your notes? 🐢

Given that today we have more convenient and sophisticated ways to capture information in many different contexts it’s very frequent that we cannot give to these things we capture as much attention as we’d like in the moment, right there.

So we send them to our future selves. But how do we get back to them? Some questions pop up here:

  • What does reviewing a note consist of?
  • How do you choose what to review?
  • How can tools help you in these processes?

Related Twitter thread

From another topic:

Do you have a rationale on what you do in place (as in, “noted and done”, only will come to it if needed) vs what you send to your future self for reviewing? Do you review notes in a consistent manner?

I have a somewhat chaotic system in place. I do ~weekly reviews but I think I have too much to review (I collect a lot) stacked up. I’m trying to find some sort of review I can do more often, or some sort of heuristic not to feel I need to think through every bit my mind chose to capture (how to choose what to review).

I don’t personally use a review system (I largely don’t review things), but I know that in the spaced-repetition software there is a sort of inversion that goes on. Rather than saying that everything must be reviewed in order at this point in time, you add things to a catalogue of things you’d like to review, and over time you are prompted to review those things (maybe more than once). You decide how much to review and how often, and the system selects an appropriate quantity of things for your review at that time. I think this, and some sort of priority could largely obviate the “having too much to review” issue.

1 Like

I can see system heuristics playing a key role in this regard.

Consider the case of important notes versus mundane notes. It is very likely that the creation and reading moments for each of these cases will embody different properties.

As a basic example, it is more likely that you would spend more active idle time while crafting important notes than mundane notes. Length, vocabulary, redundancy, they are all properties the system can use as a base to classify notes heuristically.

I think that this should be a system role: to offer affordances that will facilitate recollection of data according, not only to importance, but to any other desired property.

Yes, that’s true in theory. But I have a feeling that I’d be very frustrated about a system that decides for me whether a note is worth reviewing or not.

I think we would be surprised at how efficient constantly evolving heuristics can get. Most things we judge chaotic are statistical. A system that learns from your habits should be good enough, at the cost of little patience.

I do a weekly review, yes. Lasts between 30 minutes - 3 hours, depending on the week.

I put everything in the inbox first, I noticed that when I think about something, even if it feels exciting in the moment, usually after a couple of days it doesn’t seem to resonate anymore (so no long-term note comes from “in-place” thinking).

When an idea seems still worth mulling over during the review, I rewrite it, and put it in a “hatchery” of sorts, where I have a chance to work on it, before it hits (or not) my long-term storage.

I throw stuff out constantly, some reviews lead to no long-term storage notes.


Hi Hanbzu,

Great question.

As information passes through me, the way that I manage my review process is through a combination of:

  1. progressive summarization,
  2. merging ideas, and
  3. tags

I capture everything in a text editor today. Whenever I review something, I write a simple 1-3 sentence summary of the things that I learned from the document.

In my experience, capturing and reviewing information at the same granularity leads to situations where:

  • my review stack equals or outweighs my new stack of information
  • if pushed into something like anki for spaced repetition, the information that I could be reviewing may be so niche that it is not even worth reviewing in the first place

I prioritize what I need to review based on weighted breakdown of my tags.

I haven’t taken Tiago Forte’s class, but the table of contents for his “Building a Second Brain” he presents in his course demonstrates a great workflow for simplifying what I just described to you.

1 Like

The way that I think about this is that my personal knowledge graph (as an abstract concept, not a data structure) is constantly growing and shifting towards new or deeper areas of interest.

Tags and the relationship between tags (both distance-wise and across things I review which may be independent from one another) are how I prioritize.

It would be up to you to determine what prioritization method you apply (highest to lowest, lowest to highest, TF-IDF… etc.

This is interesting. Today I bumped into another related idea: You capture your thoughts and then put them aside, and then you come back a few months later, could you approach them from a sort of “beginner’s mind”?

That is to say, acquiring distance with the note may also help unlearn unhelpful connections or views. And, when you don’t assume you know what something is talking about you may read/reinterpret it at a higher fidelity.

This last point is specially relevant.

You don’t always have the “high fidelity mode” on. But when you’re trying to explore the territory, when you are trying to extend the map by bridging thoughts, you may want to have a high fidelity observation mode on, a sort of beginner’s mind (as opposed to one full of preconceptions and prejudices).

Could you provide an example of how you use tags in this context?

I also find powerful the idea that Tiago Forte calls progressive summarisation (which to me is very similar to what Niklas Luhmann had to do in his zettelkasten, to manage the overload of cards). What I’m wondering is, given a certain amount of things going in, what’s a good workflow that allows you to develop interesting work while you don’t spend too much time on the mundane or on notes/ideas that will not grow over time.

By the way, I also noticed that in your description you also “write a simple 1-3 sentence summary”, so you express the idea in your own words, similarly to what @szymon_k does when he “rewrites” the notes during the review (as opposed to plain clipping—or underlining if it’s a book).

1 Like


Here is an example of a passage that I took down yesterday:

A system is an entity that maintains its existence through the mutual interaction of its parts. - Systems Thinking An Operational Perspective of the Universe

From this passage, I would extract tags based on all of the nouns in the text. (The rules for identifying tags can change). There are a variety of downstream scripts I run on top of my notes in order to group. Collocation analysis is one of those methods that I use in order to group tags across documents.

It is important to note that my note taking process involves the following steps:

  1. Capturing quotes from the original document in raw form.
  2. Providing a comment, based on my own response to the quote.
  3. Summarizing all of my lessons learned into a 1-3 sentence summary for the entire document.

RE: zettelkasten - yes, exactly! The challenge I encountered with this direct approach (in both zettelkasten and Roam) is that each tag must be explicit. I want soft matching between my tags.

I have several different scenarios for reviews:

  • On a regular basis (daily, weekly, monthly, quarterly), for well defined processes; e.g. I write daily reviews collecting answers to a few questions, review a week’s worth of these every Saturday, then review weekly summaries at the end of each month, etc.
  • Learn: spaced-repetition based regular review to commit something to long-term memory. I don’t use this as much as I’d like to, because there aren’t a lot of good tools for this and they don’t integrate well into my current note taking process.
  • Work & creation: as I’m thinking about something I want to review all notes for a particular topic. That is a mix of having good index notes in place and do tagging well, plus a little bit of link exploration with a touch of serendipity.
  • Creativity & serendipity: surface forgotten notes for making unexpected connections. This is a good activity for downtime in short bursts of otherwise idle time (needs to work well on mobile devices!). I miss specific affordances for that. Something like just random note of the day, or something tied to last viewed/modified date further in the past would be a good start. I’m basically just scrolling to the end of my list of all notes ordered by last modification date and look at what I find there, making modifications so that they are no longer at the end of the list.

I’d also like to point out that although I use a tool (Bear, currently) that doesn’t distinguish between different kinds of notes — everything is a note in there, I very much distinguish between different kinds, and reviewing each kind is a slightly different process:

  • Scratchpad note: often a collection of multiple things that will later end up in separate notes or won’t be processed any further. I have something similar to the daily note in Roam. Goal of reviews is to get the valuable content out of them into regular notes.
  • Regular note: an atomic insight, trying to stand on its own. Closest to what Andy Matuschak calls “evergreen note”. Goal of reviews is to make it “better” — rewrite the content to make it clearer, add more connections to it and from it, phrase it to make context more explicit such that other people can better understand it.
  • Index note: often doesn’t include any insights itself and acts as an index into a collection of related notes. Goal of reviews is to make sure that it converges towards completeness in terms of note coverage. Lots of low hanging fruit for technology to help here. There’s some overlap with tagging here; index notes are somewhat like tags in reverse.

Tools today focus too much on content and not enough on structure, but what we really need to support our thinking is help with structure.
Something like Roam works a little better, because it breaks the main unit of notes from documents into smaller bullets, and as such makes their structure more malleable. That’s the right direction, but we’re still missing affordances for working with these smaller pieces in more flexible ways, in particular when trying to look at the big picture all these notes we take are trying to paint.