It is 2 years since I gave up on Anki and wrote @ig3/srf.
I have used it to study almost very day, for 2 years now. I have changed the scheduling algorithm several times. Sometimes little tweaks of parameters and sometimes fundamental changes to the algorithm.
The most recent big change was to eliminate the percentage of correct answers as a factor in calculating the new interval. Instead, for each review of a card with interval longer than the learning threshold, the interval and due date of every card with interval longer than the learning threshold is adjusted, according to the difference between percent correct and the target percent correct. This decreases the delay of feedback. It's early days, but it seems to be working.
Overall I am very pleased with @ig3/srf. I am limited by my poor memory but not overwhelmed and not feeling like a failure. At least, not most days. There are some days, usually after a few days without sleep, that it truly seems hopeless. But then I get some rest and get back on track.
I have put a lot of time into development @ig3/srf over the past two years, but I am confident that it was less effort for a better outcome than trying to maintain my Anki plugin to improve Anki's algorithm. I don't regret my decision at all.
No comments:
Post a Comment