Hi Dan! Yes I believe your summarization is correct. I'll add that the time costs is a big part of why context is so important. Spaced testing with feedback is definitely effective, especially when personalized to the student. Context (e.g., the type of learning materials) is indeed a big factor for how that spacing and testing should be implemented. For instance, more complex content in which answering test questions is more time consuming (like Anatomy and Physiology) may mean that the relative speed advantage between correct and incorrect I describe is lessened and that practicing at higher difficulty is more efficient (as opposed to fairly low difficulty being more efficient with word pair learning). So spacing would still be good, but the amount spacing spacing that is optimal (and thus how much difficulty is imposed) would be different depending on context.

As for recommending things like Super Memo, they are still typically better than alternatives. People have a hard time scheduling their own practice efficiently, and SuperMemo helps do that for you. So I imagine it is still superior to a fixed schedule of practice (e.g., repeat every X minutes) or ad hoc methods a student may do for themselves (e.g., sort items into different piles a la Leitner method). I don't believe SuperMemo considers time cost differences between correct and incorrect answers however, so it is probably suboptimal. Please reach out if you would like to chat more!