How to take notes: not self-help advice

I think I have a unique take on note taking.

"Notes" on literature etc. are different than hard sciences, see below. I'm not a fan of 21st century self-help note-taking methods like mind-mapping or playing with different colour pens.

One form of note taking is to use your knowledge to write up a little essay on a topic in order to figure some things out. This is something I learned from Paul Graham's essays. You can see them here. Here's a useful one: How to Write Usefully" I think these can be quite an asset for understanding different topics. This stretches the definition of "notes", because you can start writing these at any time on any topic, so they are often not notes taken while consuming some material, but fresh new work from daily musings. I find that this works best when you start out not knowing the answer to the thing you want to write about. Makes for an enjoyable writing experience. I'd consider this article to be such an example: I didn't know what I was going to say about note taking until I started.

Otherwise, one can also be solving some math problems, mashing out some pseudo-code. Whatever works. This is something I've learned from Dijkstra, primarily. Perhaps also the tendency to write a DSL that I learned from SICP has led to me writing arbitrary pseudo-code made up on the fly in notebooks. This is what I recommend for "hard sciences". Anything where you're doing math, drawing diagrams, etc. Interspersing with English commentary is helpful. This is modelled on Dijkstra's exact style for discussing the fundamentals of computer science.

I've found the best way to do this is to have a single notebook where everything goes into it, a notebook that I carry around with me and sometimes leave open on the table so I can easily start writing into it. When it's full, I number the pages and make a table of contents, because otherwise it would be hard to find anything in it. Separating things into separate notebooks doesn't work for me. I never have the right notebook with me, I always forget which notebook is which etc. That worked for college when I'd have a separate notebook for each course, but for auto-didactic stuff there is no course. I prefer a notebook without lines, for the math and diagrams, if I can get one. Just like Dijkstra.

I've scanned a few thousand pages of these kind of notes with my phone, and plan to turn some of them into typed-out and redacted digital versions for my own future benefit. Some I might publish on a blog. Most are garbage!

Guarino Guarini, sometimes seen as the first renaissance teacher of Latin, suggested taking little quotes out of books and putting them into a notebook for later. I do this now and use grep to find stuff in that notes folder. Thanks to the joys of hoarding books in Calibre, (acquired from the LibGen of course) I know I will be able to search up any portion of these quotes in the ePub I downloaded for instant referencing, as needed.

Guarini himself (from Breaking Tradition)

• If then, in the course of the reading, you happen to discover a nice saying, a wise and noble action, a quick answer, something pertinent to a gracious way of living, it seems to me that you would do well to keep at hand a little notebook like a faithful depository in which you might make notes of points in the text; in this way you will be able to go back to your favourite parts without having to leaf through everything. […] But if copying things in the notebook causes annoyance and interrupts your reading, you should [not do it].

I find that this is a good way to take notes on non-techincal books. Literature, self-help, business books, whatever. First I write them on any paper I have around, then I select from them the best ones to type into the computer. This gets you a filtered version of the book that is the most important, meaningful, and interesting to you. You can also show off by pulling out quotes like the one above. The Guarini method is nice because it gives you a superior reference for the future, better than referencing the book itself. In the time of Guarini (13th century), it was also noted to be better than owning and carrying around dozens of heavy and expensive books.

For books on computing or math I prefer to read a chapter, then grind my gears and try to figure something out that I find interesting. Often times it is something a bit goofy that I come up with, like how to derive sin(x) from cos(x) or whatever. Or it might be related to another software project I'm working on, and I'll decide to try rewriting some bit using this new method I've learned. If the book isn't hard enough to enable this kind of reading, it probably is too simple of a book.

There's no reason to take notes on books the way you take notes on lectures: you can easily reference books, but not lectures. Even if you are consuming video lectures, they just aren't easily referenced like a book. You take notes so you can have a lecture with pages, it seems. The goal isn't to memorize your notes, it's to have something to reference while you do the homework. I noticed in college that often times every student's notes were virtually exactly the same for a given class, from philosophy to math to physics, yet I felt that it was necessary to make those notes anyway. Then I could easily flip the book open and look at the notes for a given topic while I did the homework. Again, I find that watching video lectures is pretty useless unless there is some kind of homework I can do. If the book is good enough, the lectures feel pointlessly redundant too.

I'll finish with another quote from my collection by Guarini:

• Do you desire outstanding advisors, the best, that is, in terms of faithfulness, beneficience, prodence? Then search out books. Read them. Make it a habit to carry them with you always, in the countryside and while travelling, as others do with dogs, falcons, and dice. I would prefer you have more books than clothes. Not only are your books more useful and pleasant, they are also a good investment. You'll always be able to sell or exchange books, but one can't say the same about cloting.

Another person worth referencing would be Dijkstra, who took notes with his fancy pen, tried to make them readable on the first try, and faxed them to his colleagues. You can learn a bit about this in the EWD EWD1300 , but I think there was another EWD which better explained his unique method.

I've read dozens of EWDs, but I can never reference them, because I didn't know about Guarini's method at the time!

To summarize: for literature etc. I use the Guarini method to create a valuable asset supported by centuries of studies in the humanities. For math, I try to take notes similar to how Dijkstra did it in his excellent EWDs. For actual code, I prefer to make up a language for the problem on the spot, and start using it on the paper in order to solve the problem. And for figuring out answers to non-specific questions like "what is a good way to take notes", writing an essay on it is the way to go.

Last week a friend of mine asked me to "start documenting stuff". This is because I unload erudite wisdom from a week of reading when he comes around, just talking to him about it, and the information from that discussion goes up into the air without anything to show for all of that learning and thinking. There's papers all scattered around the office. Correctly, I pointed out, that if I documented stuff for him, instead of just for me, at that point it would be just as easy to document it for everyone. Here is today's documentation.