On Notetaking

Contents

Analog

What is the most efficient way your brain has learned to store and connect knowledge? For most of human history, people used handwriting to record their thoughts.123 Even kids starting school since the year 2000 still learn to learn by handwriting notes. I certainly learned to learn this way, and I don’t undervalue that as a factor in my ability to learn a topic. Some disciplines (e.g., electrical engineering) lend themselves especially well to having a page on which to draw diagrams.

The wrong way

I’ve always held a great deal of value in taking good notes.

On this count I made two critical failures when I started college: failing to date my notes and writing them on loose paper. The first is a problem because the more you write, the less you’re able to remember exactly when you wrote something. The second exacerbates the first.

If you take notes, write dates on them. It’s a small effort that solves a whole class of problems that should never exist in the first place. This is silly, you might think. If I take my notes in a notebook, like a regular person, I won’t even have to date them. Don’t succumb to this trap. Just write the dates.

I used to use a clipboard and printer paper, or a clipboard and graph paper. I didn’t like spiral notebooks because I’m left handed. Once you write enough on loose paper, it eventually gets mixed up. You’re left with a pile of isolated flakes of knowledge: like newspaper clippings if someone blotted out all the dates.

Choosing the right tool for the job

Legal pads are awesome for left-handed people. Your hand won’t be crippled against a harsh metal spiral, and the left and right sides of a page on a legal pad don’t matter, since the spine isn’t vertical anyway.

I write my notes using a modified version of Cornell notes that don’t include the part where you write a summary at the bottom of the page. The point of doing Cornell notes is to provide a skimmable index for your notes. When you go back looking for something you wrote a month ago, you want to be able to find it quickly. Legal pads already have the margin for this, but I like wide margin ones that allow for more detailed margin notes. If you want to go back and add footnotes to something you wrote earlier, getting the wide margins pays off. One Amazon purchaser wrote:

First used these in law school and got hooked. The wider left margin is great for jotting down specific reference points and the right side for a more detailed explanation.

The downside to legal pads is they are designed to be abused. The pages are usually perforated and if you carry one around for long enough they will fall out. Further, the paper degrades over time—if you keep one on a shelf for a decade, it will fall apart when you so much as bump it.

I spent a lot of time looking for the right notebook to use instead. There are a couple important qualities to look for in notebooks:

  • Binding. The pages should be bound. This means stitched together with an actual piece of thread. Spiral notebooks will rip your paper over time, and binders will do likewise (and require various horrible remedies over time, like those hole punch reinforcements you saw in elementary school). Legal pads will come loose at the perforation.
  • Paper quality. Paper should take well to markings and not smear ink nor pencil. Even if you use a fountain pen, ink should not bleed through. It should be acid-free so that it does not yellow and degrade over time.

Options that meet these criteria are everywhere. The best value I could find was Black n’ Red Casebound notebooks. There is no indication of whether the paper is acid-free, but they are high-quality and anecdotal evidence from A. Shrestha suggests they might be:

I have one that is nearly 10 years old, and I don’t see any issues with yellowing, fading, or deteriorating.

On each set of pages I draw margins on the left/right sides, respectively, so the page looks something like this:

    +--+-------+--------+--+
    |..|,...,, | ,..,,. |  |
    |  |,..    | ,.,    |..|
    |..|,.....,| ,...,. |  |
    |  |,..,,  | ,...   |.,|
    |  |,...., | .,.... |  |
    +--+-------+--------+--+

If I have some spare time before a class starts, I will draw margins on more pages and write numbers in the upper corners. This helps prevent slowdown and misaligned edges when you don’t have time to fumble with a ruler, and gives you a system for referencing earlier pages so you can tie the individual bits of information together.

Digital

Beyond the neon shudder of Ninsei, the sky was that mean shade of gray.4

When I first discovered LaTeX it took me half an hour to learn enough to type my calculus notes. When I found a symbol that I didn’t understand yet, I would type an identifier for it and go back to define the command later. This was slow, but eventually I got the hang of things: once you learn a symbol, you only need to look at your old notes to figure out how to type it again.

I have three good reasons to type my notes:

Memory

I used to think I was bad at memorizing things. It turns out I just didn’t know the right tricks.

I started using Anki seriously in September 2014, during my senior year in college.6 I use it to memorize chemical formulas, programming language syntax and libraries, Linux syscalls, and other similar facts. Anki is spaced repetition software, which means it exploits the psychological trick that the best way to memorize something is to review the material right before you would otherwise forget it. Each time you do this, it extends the duration that you can keep some tidbit of information in memory.


  1. Every college student has encountered a professor who won’t allow students to use laptops in his/her class.

  2. Laptop use can negatively affect performance on educational assessments, even, or perhaps especially, when the computer is used for its intended function of easier note-taking. Although more notes are beneficial, at least to a point, if the notes are taken indiscriminately or by mindlessly transcribing content, as is more likely the case on a laptop than when notes are taken longhand, the benefit disappears.

  3. William Gibson, Neuromancer, pg. 15.

  4. According to a totally unscientific test that I just took, I type at 92 words per minute (WPM). While my theoretical maximum WPM is probably higher, the I’m not trying that hard speed is much closer to what I get when I’m taking notes and not, for example, transcribing text.

  5. Some people suggest using pre-built decks. I’m against this. One Less Wrong user agrees with me:

    I’ve been working my way through CLRS lately, and I notice that I’m able that extracting all of the important content in a chapter and transforming it into cards takes maybe about 30-60 minutes. Previously I would spend much longer on the same material, and comprehend it much worse.