Learning: lazy loading and other methods

[badlydrawnrob]

Time tracking

So far it's been at least 4 solid days on the chapter, plus to-do notes, add cards etc.

What is "lazy loading"

The Janki Method is a great way to learn, and for learning a bunch of syntax it can work well (if a little tiring adding the cards).

I've found lately that a good way to build cards (when reading a book, say) is to add learning points as you go, group them into related cards, reducing those learning points down to the bear minimum. Many of the notes taken during a chapter are best as a single file you can reference, with key points fleshed out as Anki cards.

I call this "lazy loading" of cards. It seems to work and results in fewer cards to revise.

Might be wise to close some of these in the Anki repo, and create proper articles with my thoughts, once I've collected them and done a little research.

Revision can take a long time

Could simply trying to code up chapter code a few times give me the same results?
You'd still have to find some way to memorise the way the code links together.
Perhaps there's no easy way to do this (learn heavy functional stuff).

For example, Elm-in-action chapter 03 took me a 2-3 days to get through while making lots of notes and adding to my to-create-cards list. After that a couple of hours sifting through the list and grouping related content.

The list became quite long, and it might be easier to print it out to see the links between things. Then using the circle, underline, and make notes technique you could smush that list down further.

I've still had to create some standalone (simplified) bits of code to link through from my cards. I honestly don't know the best/quickest way to do this. It seems to just take time.

The problem with this method

1. You can end up with a big laundry list of to-dos
2. Highlighting and bookmarking of the eBook becomes poorly maintained.¹
3. Maybe I shouldn't have to re-read the book again to get the jist of the ideas it contains.²
4. Reading code from the book perhaps isn't as easy as well-formatted, coloured, code blocks³

The problem with large code bases

An Anki card becomes harder to create when the "problem" space is across different lines of code. You have to simplify it somehow, but if you don't have all the information needed to understand the problem will be available.

• This bug would be easy enough to simplify and show
• Refactoring this chunk of code might not be so easy to ask a question about.
• Even a small amount of code; being able to reason about how it fits together isn't so straight forward.
  • You might get away with the "jist" of the code in shorthand form
  • Or adding just a few lines with "imaginary functions" (ones that aren't written down)
  • Then link through to the proper code chunk.

Options

1. Add a link to the whole chunk of code with a simple question on the card
   • You'd have to look at the external file and answer the question
2. Reduce the chunk of code down to simple comments in the form of "steps"
   • Kind of like Dr.Racket's "stepper" ... simple black boxes with inputs and outputs
   • You can succinctly describe the problem and only show the core bit of code.

Alternatives to cards

1. Just use the compiler! Elm's compiler is very good at notifying you of errors.
2. Is the documentation good? Just use that then!
3. Has somebody else made the notes that you need?⁴
4. Are there any Anki decks available to use?⁴

Other mental models

• Getting things done
• Barbara Oakley's books (like "Learning to Learn")

Relevant issues to mush together at some stage (into articles):

• Learning: A better way to create cards? #10
• Draw! anki#50
• One idea per slide? anki#41
• But I'm a kid! I don't know that yet. anki#103
• 3 times, 3 ways anki#93
• Trying a "lazy loading" style of adding cards anki#91
• How to learn coding in the age of Artificial Intelligence? anki#92
• Using the Anki app (the good and the bad) anki#72
• An alternative to puzzle? anki#40
• Anki note dumps anki#89

Footnotes

1. With the MacOS Books app, you end up with a ton of bookmarks and highlighted stuff. I've tried to prioritise with different highlight colours, but looking through the book can seem like a rainbow of disorganised notes. Grouping together of related ideas becomes very important the more there is to learn. Ditto condensing and being concise about those ideas. Learning terminology has never been my strong point, but a few well-picked Anki card examples might be better than a chapter full of highlights. ↩

2. Creating your own notes file per chapter, some well-picked examples, some links to documentation, perhaps even your own documentation might be good enough so that you don't have to revisit the book again. Smush it down to get the jist of the idea, per chapter. Finding the exact thing you need to re-learn within the book isn't as easy as following the book step-by-step. ↩

3. Such as this example. Or this one. The book is very good for step-by-step instruction and interleaving of knowledge, but might not be the best reference source. Are you having problems remembering something in particular? Write it down! Create a clear example! Find someone else's code who's done it for you already! ↩

4. It's generally better for learning if you create your own notes, cards, examples. They'll probably stick in memory better. However, it's better to cheat sometimes, too 😉 ↩ ↩²