Just a general recommendation: for math/physics/CS it's far more useful to work through problems than memorize facts. I'll use linear algebra as an example since that shows up often in your cards.

Say you want to learn about eigenvectors and eigenvalues. Instead of writing the definition, you might have a card which asks the following:

"Create a matrix whose eigenvectors or eigenvalues are always equal."

This will force you to do two things. First, you will learn from creating the card since you're forced to synthesize the information rather than copy. Second, your recall will force you to work through this newly created problem.

Another general strategy is to just pick a problem from a book that uses the concept. If you ever get the problem wrong, you oblige yourself to solving an additional related problem from the same book (i.e. find the basis of one of the following matrices).

I guarantee you will learn 10x more effectively.

Also beware of falling victim to spending more time on building tools than learning... I've found that pencil and paper go a lot farther than people give credit.

I don't think it's one or the other. I struggled a lot through differential equations because I didn't have good memorization skills. Had I memorized all of the formula it would have helped me get more out of doing the problems.

It's like learning Spanish. If you memorize a bunch of words you won't be able to speak Spanish. But if you have a bunch of Spanish words memorized than you'll get a lot more out of reading/listening/ trying to speak Spanish than if you don't.

And memorization is usually a far less taxing activity that you can do on your phone when waiting in line or using the bathroom.

there is some truth to this but some areas of math do really benefit from just raw memorization. At a certain point, remembering the definitions and the statements of theorems may not be so trivial, and having them in your head makes it much easier to solve problems.

I agree here. Actually, in most cases, it was enough to know only the definitions. I could come up with applications from definitions on the spot. The thing to watch out for is to get used to the way to solve problems and forget the definitions. I think this is a more frequent occurrence (especially when preparing for the exam.) Then the one will get trouble if they need other applications.

Yup. There are a lot of physics concepts I broadly understand, but solving a problem for them is a different beast.

edit: Also, speaking as someone who mostly taught themselves physics from the books in college and didn't go to class. (I was not a great student, I did _okay_, and probably got lucky).

The caveat to this: I would recommend putting such questions in a separate deck. Solving these problems take time.

What I do: Fact based stuff that I want to just recall quickly go in one deck. I do this deck daily - 5-10 minutes max while having breakfast.

The more complicated cards go in the second deck, which I do only occasionally (e.g. weekends when I have time).

When I initially had them all in one deck, there was no way to do a daily review in a few minutes, and so I would often do it less frequently which messes up the whole point of SRS.

It seems to be a good way to assist with main usage. But I think it's dangerous to try to compose questions with it alone. In my personal case, I found it too burdensome to create such questions. Of course, if I want to learn something really deep, the way you said would be good.

I like that you allow testing the app without creating an account first. That's good UX.

Thank you. I also like the idea of having a separate Playground and making it like a pixel-game.

I wanted this at university and tried to write something similar before inevitable employment and distraction. I wish you every success, a couple of ideas that may be applicable:

Probably quite cheap to pay people to onboard sets of well established notes, eg cambridge undergrad maths notes or similar. This may help with dedicated users in niche communities.

Collaboration is hard, but a forking mechanism might be easy+valuable.

Tree structure could guide spatial repetition probabilities, see eg that half-life regression paper from Duolingo.

Thanks for the suggestion. The app already has export and import functions, so it would be nice to have established notes. But I also think people need a note-taking and organizing something by themselves for learning. I think that's why people take notes even though they have well-organized tons of books.

I'm not really sure what it means for the forking mechanism. I would be grateful if you could show me a link to the resource for the forking mechanism you mentioned.

As for the tree structure, the tree structure is already reflected to some extent. As with breadth-first search, the questions starts from the higher level, but when the parent node is in the wrong state, the child node does not appear in the problem. I'll search more about the half-life regression you mentioned!

People would publish flashcard sets (share part of their notes tree) and others could fork and modify those sets, like GitHub. This is in my plans if I ever have enough time to compete in this space.

I think if you pay not much money for this, you'll get bad results. It's very easy to make memorisation materials, but it's hard to make good memorisation materials! Your best bet might be to convince a few Cambridge students that their path to victory lies in making these resources for themselves, and then pay them to give the resources to you afterwards.

The idea is awesome. This is something I didn't know I needed, the biggest advantage is the hierarchical structure makes anki cards much easier to manage. I'm totally your target audience. I like Anki but find the interface for making cards cumbersome.

But a couple of pieces of advice.

1. I think there is too much stuff on the screen and it can feel overwhelming.

2. It doesn't seem obvious how to use it. I think if the first screen you went to was a hierarchical example deck it would useful. An illustrative example is worth a 1000 tool tips.

1. Okay. I got that there is still too much stuff in the app.

2. I thought the Playground is for those things. Was that not enough?

Yeah it's there, I just think it's hard to find. If I'm new to the app I won't immediately click playground then example. My first time using the app it took me a little while find it. I don't know if this was other people's experience, but here is what mine was.

I went to the website and logged in. I'm trying to imagine how I would use this application instead of Anki, how do I enter/organize information and how do I memorize it.

I'm greeted by a blank screen. Ok I've used workflowy before I see that I can generate some of type of hierarchical structure. But how would that information turn into cards? Ok let me click memorizing. It says I don't have anything to memorize.

Let me try to find an example so I can see how the cards works. Wow there is a lot on the screen, let me try clicking different things. Lots of clicks later... Ahh here's an example.

Clicked memorize. Oh that doesn't work, it doesn't apply to the example. Let me try to copy this over. Oh I can't. Ok looking at this example I know how to build out the relevant structure to generate the cards I want. Build out structure. Ok now I click memorize. SUCCESS!!

I think if I landed on deck, then I'd immediately see how to organize information. And when I clicked memorize it would work and show me how the organized information translates into cards. It greatly shorten the "arrival -> AHA" time.

Okay... Thank you for telling me about your experience. I got where the confusing point is.

Maybe you've noticed it yourself now, Playground is only for organizing parts. I think it would be nice to have a Playground for the Memorizing part, but now it doesn't have that one... To make it appear there you have to put the content in the real Organizing part. Anyway, I found out that I need to make that part a little more understandable.