Spaced repetition idea

[hamoid]

First, thank you to @digitapex and @machiav3lli for this app! Super useful!! :)

In the description of the app I read that after answering 5 times correctly, a word is never shown again.

If that's true I think this is not ideal. If I spend half hour learning, I can get words right 5 times in a row, but that doesn't mean I'll remember them one day or one week later. What do you think about adding further repetitions with exponential delays? for example 1 day later, then 2 days later, 4, 8...

I don't know what's the ideal progression. I found this page talking about it: https://e-student.org/spaced-repetition/

I think adding these extra repetitions would make sure that I actually didn't have things in my short term memory.

[hamoid]

Maybe a simple approach might be that for every mistake the number of expected right answers for this word is increased? So if I answer 5 times right and never make a mistake it means I know that word. But if I I make one mistake, I need to answer it right 6 times (with spaced repetition).

With some words I insist on answering wrong, so I think the program should really make sure I have rewired my brain :)

This would involve replacing the current constant (5) for a value per word. I have only looked into the code a bit. Are all user actions stored in a database? That would be good to figure out which words the user struggled with (if right/wrong counters were kept per word).

[machiav3lli]

Hey, I'm already aware of spaced repetition and have considered implementing it. But for now, my other (FOSS-)projects and private made the project on hold (or in maintenance mode if you'd say). But maybe in three months I'll have some time to invest in some improvements.

[mvhulten]

The idea that words come back again after some time seems like a good idea at first. However, how it works now works well for me so far. Even in the future, when I have likely forgotten some of the noun genders, it is okay if those words are not repeated. I already get those repetitions through other ways I interact with the German language (other software, usage in real life, classes, reading, ...).

If such spaced repetition is to be implemented, the algorithm should be chosen in such a way that learners do not get annoyed that words are repeating (e.g. if they haven't touched DerDieDas for a month but have used many of the words already learnt—at least, still, somewhere stored in their brain—I would prefer that I get new words). An increased correct-answer repetition could also become annoying.