Postscript: After One Year of Mnemosyne

Added February 4, 2009

At the time I first wrote the Mnemosyne/SuperMemo/Anki review, I intended to return to the programs one year later and reevaluate them. After a year of using Mnemosyne, however, I have decided not to do that. Mnemosyne meets all my needs, and has given me no trouble in the year I have used it. There are very few software programs I can say that about.

Also, I've contributed too much to the Mnemosyne project at this point; I think it would be unfair for me to review the software against other programs, as I have been supporting Mnemosyne since shortly after I reviewed it for the first time.

Mnemosyne hasn't changed too much in the previous year, although a major new version, 2.0, is in the works. But, aside from scrollbars, undo, and autosave, there are no features I wish Mnemosyne had now that it doesn't already have. The software is finished, which in the Google world of “everything beta” comes as a great relief to me.

To me, Mnemosyne's stability and reliability are what's most important. I use the software for about two hours every day; even an occasional crash or bug could mean an enormous amount of work lost. I had many such problems with SuperMemo. I have not had such problems with Mnemosyne.

As for the changes Mnemosyne has undergone, two are especially notable. First, there is now a system of plug-ins for the software, allowing you to change the basic functionality if you need or want to. I use two plugins, one that changes the color of the cards and another that increases the randomness. I like the plug-ins system especially because it means that a cool new (real meaning: untested) feature is not able to introduce bugs into the main program.

Second, the Mnemosyne website now hosts downloadable card collections, of which there are several study sets for Japanese. I don't recommend using these-- I think you ought to make your own cards-- but for those of you who want them, they are there. What I have found useful are the geography cards, and not just for Japan. For memorizing the names of places, an individualized approach is not that beneficial, and making maps is time consuming, so I'm grateful to the many people who contributed to the collection.

In the end, the thing I like most about Mnemosyne is its simplicity-- there aren't a lot of features to get in the way (unlike SuperMemo), and I appreciate that now even more than I did when I first made the switch. Curiously, I find I don't miss most of the advanced features I used to use on SuperMemo, including even incremental reading. I still think incremental reading has some amazing uses, but I've realized acquisition of foreign languages, particularly Japanese, is not one of them.

I've used Mnemosyne every day since I first started more than a year ago now. And in all that time, the most annoying problem I have encountered is that I lost about thirty minutes of work because there is no autosave feature and my computer crashed before I saved the data myself. Considering that I've used the software for more than 700 hours at this point, I think this is the greatest testament to the program's usefulness.

The algorithm of Mnemosyne (and other open source software like it) is far more basic than the one used in SuperMemo. And to be honest, I like the “feeling” of the way SuperMemo schedules repetitions better. It's more random, and yet more familiar-- or so it felt to me. But, having used SuperMemo for eight months, and Mnemosyne for one year, in my experience it appears that in the end both algorithms work about the same in practice, which means there's no need to deal with the bugs and quirks of SuperMemo.