[Smaug123]

Tip for life: human memory is essentially a solved problem. Spaced repetition Just Works (TM). Learn how to use Anki, read Michael Nielsen on Anki, read through the first essay of https://quantum.country (even if you don't intend ever actually learning any quantum computation) just to see what good spaced-repetition flashcards look like and how to convert a text into flashcards. You can memorise very nearly anything with Anki.

Edit to add: even if for some unaccountable reason you choose not to use Anki or one of its siblings, you must read and comprehend https://www.supermemo.com/en/archives1990-2015/articles/20ru... .

[klenwell]

Good tip, but I thought you were going to go a different direction with this. More like:

- Don't bother memorizing trivia. At least for work. Save room for the pleasurable trivia from your private life. Convince your team to set up a knowledge base and use it.

- Learn how to bookmark and reference stuff that is important.

- Almost nothing said in a meeting matters unless it is written down and most of the stuff that happens in meetings immediately evaporates into thin air.

- The palest ink is more reliable than the most powerful memory.

[8589934591]

Are there examples/blogs of how to use Spaced repetition for math/sciences/computer science/programming?

For discrete math currently, I solve problems, and either I come across 1 complex problem for each topic which I might have struggled with, or a problem which has concepts of multiple topics interleaved. I add that to my deck and I solve them regularly.

[Smaug123]

Nielsen has http://cognitivemedium.com/srs-mathematics, for example.

I would strongly recommend not having entire problems as a single card; it's absolutely crucial that answering any particular card is a single mental motion. (You can get by with a very small workload if your cards are more complex than that, but it doesn't scale even in the medium term.) I wouldn't go any more complex than "what is 75 * 43".

As an example, the definition of a red-black tree for me is six cards:

* What, imprecisely, are the colouring rules on a red-black tree? ["global; local; base-case"]

* What are the leaves of a red-black tree? ["null"]

* What is the base-case colouring rule of a red-black tree? ["leaves are black"]

* What is the local colouring rule of a red-black tree? ["red node => black children"]

* What is the global colouring rule of a red-black tree? ["black depth is well defined"]

* What is the black depth of a node in a red-black tree? ["the number of black nodes encountered on a path from that node down to a leaf"]

One might naively have created a single card that is "what is a red-black tree?", but in my experience such a card is too big. The difficulty of learning a card grows at least quadratically with the number of mental motions it takes to answer that card, and small/simple/easily-learned cards are incredibly cheap in that Anki very quickly learns not to show them to you if they're genuinely easy.

Maths-wise, my cards usually look like:

* "Does path-connectedness imply connectedness?" [yes]

* "Does connectedness imply path-connectedness?" [no]

* "Counterexample to connectedness-implies-path-connectedness" [topologist's sine curve]

* "Definition of the topologist's sine curve" [union of the y-axis and a squashed sine]

* "Definition of the squashing of the sine component of the topologist's sine curve" [whatever]

* "Main idea of the proof that path-connectedness implies connectedness" [use "all locally constant functions into [0, 1] are constant"]