speak chinese like a native

Declaring SRS Bankruptcy

Posted by SF_Rachel August 13, 2012 in the Group General Discussion .

I’ve been using SRS via Anki / Ankidroid for a couple of years now; it’s an integral part of my daily study routine. Even on days that I don’t study, I plug away at my flashcards. It really helps with momentum. Over time though, I’ve become concerned about the growing size of my card deck; Anki doesn’t really seem to want to handle large data files (for instance, the online service only supports decks < 10MB; mine is now 3X or 4X that). Every week I see increasing hints that catastrophic data corruption is in my future.


Now, I’ve long suspected that the data size of my deck was due to the possibility Anki retains way more review data than is really necessary. That is, most of the data in my deck file is the review history of each card, not the card data itself. Anki retains each card’s entire lifetime history. It seems to me that the most recent 3 reviews + most recent single interval ought to provide enough predictive power to calculate a useful next interval. I have no idea how other SRS platforms handle this, but I do suspect Anki’s is not the most elegant or data efficient.


That aside, I really do love Anki, and have finally even gotten it configured so it displays cards exactly the way I want (it displays the hanzi really big, the pinyin is there but in a tiny font that is hard to see. That way my eye doesn’t go right to the pinyin, while I have the safety net of knowing it’s there). I hate to abandon my deck too – I’m a little bit OCD and certainly a completist, and I do love that Anki has some nice data analysis features that let me see things like predicting my card load for the future, counting my Hanzi, etc. Nevertheless, I think I’m on the verge of declaring SRS bankruptcy and starting over on a new platform.


Comments (15) RSS

loading... Updating ...

New lesson idea? Please contact us.