[tedks]

>The hardest and most time consuming part of using standard SRS software is building your deck.

This is because that's the initial learning phase. You can't just download a spaced repetition deck and run it though any program and learn it in any meaningful sense. The cards I need to create to learn a concept will be totally different than the cards you create and so on.

Connecting your cards to the relevant nodes in your personal semantic network isn't just the most effective way to learn, it's the only way human brains encode information. If you're not doing this purposefully, you're just trying to glue someone else's relevant retrieval cues into your own semantic network.

There are other ways the comment parent's site could avoid this, but inevitably in my experience this leads to FALSE learning, where you have very sparse retrieval cues, basically limited to what you see on the screen in the learning app. This gives you a feeling of progress, because you really can recall the information in the app, but very little practical use.

[sheepmullet]

> This is because that's the initial learning phase.

> The cards I need to create to learn a concept will be totally different than the cards you create and so on.

I don't use SRS software to learn a concept. I use it to remember concepts I already know. As such creating the cards is just tedium.

For example, I used to do a lot of 3-6 month contracts, and one contract might be all backend Java, and then the next might be all Javascript, and then the third might be in Rails + Javascript.

SRS software allowed me to stay "fresh", at an advanced level, on around 6-8 languages and platforms despite not using them for long periods of time.

[tedks]

Ugh. s/learn/remember/ for your use case then. The cards you need to remember a concept will be totally different than the cards I need to remember a concept and so on. Creating the cards isn't "tedium", it's reinforcing the concepts you decide you need to be reinforced while presumably not things that you definitely know and will never forget because you initially overlearned (this is a technical term) them.

[sheepmullet]

> Ugh. s/learn/remember/ for your use case then.

Learning and remembering are very different processes.

I've had great success sharing cards and decks and haven't found my own cards any more valuable than the cards I've gotten from friends.

[tedks]

I think it's disingenuous because my point is the same.

If you haven't found your own cards more valuable than cards other people made, you're making cards wrong, and could be remembering them MUCH easier.

[sheepmullet]

How do I learn to create cards the right way?

[tedks]

The way you remember is you take your recall context, extract all the recall cues, and associate out from those concepts until you hit the target memory. Obviously this is a gross oversimplification, but it's basically a graph breadth first search. To make things easy to remember, you have to link them to as many other nodes as possible, and you have to link them to nodes that are strongly linked to the rest of your semantic network.

So, the best cards you can make are cards that refer to concepts, internal to you, that are highly associated. What I try to do is encapsulate the statement that originally made me understand the idea, which is usually a metaphor or analogy for another concept I've deeply overlearned. Then I make a few other cards that elaborate on the concept to get away from the metaphorical link and into the specifics.

See also: https://www.supermemo.com/articles/20rules.htm . The rules that are related to what I describe here are 11-14. This is written by the godfather of spaced repetition.