One core part of the system is the repetition spacing algorithm
that’s used to determine the next time an item should be rewiewed,
I’m using a mix of the original SM2 algorithm (developed by Piotr Wozniak 30 years ago)
and the algorithm used by Anki (which is also based on SM2).
The main difference between my system and Anki is that I’m building a SRS extension for a text editor (Emacs) while Anki is a SRS with a (limited) text editor in it.
This allows me (and the few other users) to write “full” documents, marking only some parts as flashcards.
A very interesting application of this is “Incremental Reading”, importing a document and reading it in small increments while editing down each paragraph / part until it’s gets small enough to be turned into a flashcard and memorized.
In theory, it would be possible to read / learn many documents at the same time,
switching between them once one ceases to be interesting.
In practise, importing documents / books into suitable format is tricky.
Example: I’m learning for a bunch of exams at the moment and I’ve copied all the slides to org mode,
marking all sections as “TODO”. Each day, I pick out one section and read through it, either splitting it up into smaller sections (again marked as “TODO”) or turning sections into flashcards once I’m sure I’ve understood what’s going on.
As far as I know, a similar workflow would be impossible with Anki without using an external tool.