I’m really glad Anki is working well for you. Wenlin is certainly the best software tool out there (despite the ancient look) for Chinese language learners to work with. I have been a faithful user of Wenlin since 1999.

I see that the program meets your needs, which is fine. Keep in mind that Anki has gotten the highest praise of any review so far so my criticism is in the hope that it will continue to improve in future. However, I would venture to say that you are probably in the minority in terms of the way you do things. Think of the thousands of students in language programs around the world who are learning out of textbooks who are inputting terms directly from a glossary, etc. In my case, I type them directly from books I read in Korean, Chinese, or Japanese into my vocab software. In the case of Chinese, I often copy and paste from Wenlin, but when a student inputs from a glossary they will probably type directly from a textbook (there are numerous ways of inputting tones without the numbers).

You (and I) may not care about grades and tests because we are not affected by them, but a great many students, if not the vast majority, are learning words as they go along in a class and must also think about their immediate study needs. I strongly believe a good flashcard program should make an effort to meet their needs, or accept its place as meeting a small niche of the potential demand for the product.

I fully agree spaced repetition is a must and should be the primary mode of study (I have done it almost every day for most of the past decade, in three languages), which is why I practically dismiss any software that fails to include it. However, for broader appeal, a flashcard program really should not neglect the vast majority of students who are in classes/programs where short-term needs for review on demand are also required.

Cycle elimination is a must and one of the big things missing from Anki. When you go through, say, 50 cards, and you had forgotten 10 of those cards - do you remember all ten of those cards after being shown the answer just once? If so, you have a remarkable memory. Even after being shown the answer, I will forget many of them, especially if there are many forgotten words out of the total reviewed, and need to continue “cycling” through the cards I got incorrect until I remember them all. You aren’t being shown a card over and over again immediately after, but each time you have “cycled through” all the remaining incorrect cards. This is borderline common sense and almost all flashcard programs (even the worst) do this. I’m really shocked at this absence in Anki.

Think about how you memorize a group of paper flashcards: you cycle through the cards practicing them and, as you memorize them, you put the card to one side. Eventually you have put all the cards aside and you remember them all. That is cycle elimination. You eliminate known cards each time you cycle through, repeating only unknown cards. I don’t want to wait 10 minutes! I want to learn an unkown card now, not in 10 minutes! How many times am I supposed to boot up my flashcard program? If I don’t remember it now, why should I remember it in 10 minutes?

Anki has some great things going for it and I’m very happy it works for you, but it will not meet the needs of a very large number of language students without cycle elimination and study on demand.

A nice coincidence to run into you at Starbucks. I am so glad to have found a site like this– definitely an important source of wisdom for language learners. Anyway, thought I would chip in my two cents:

I started using Anki 3 months ago, and now have input roughly ~2000 facts (4000 cards). I have never used any other software. The vast majority (~80%) of these facts were not added in order to prepare for a quiz or test. Rather, they were words I ran into during conversation with my Chinese roomate or friends, words I came across in literature, or words I saw in movies/on television. In fact, I often neglected classwork in order to review this ‘random’ assortment of vocabulary. For learners who attach a great deal of importance to the grades they earn while studying, this method is less than ideal. However, I personally believe that grades are unworthy of emphasis, particularly when studying a language. I have faith that with abundant enthusiasm and dedication, mastery will come naturally (and then so will grades). For me, it is difficult to conjure up said enthusiasm when confronting the words which the creator of my class textbook happened to decide I should study. Which isn’t to say that I don’t study them. I do– but just not with any special emphasis. And so I find no need to have study on demand. I believe that all the words in my database are equally important. Also, spaced repetition has thus far proven to be an excellent and reliable method for learning all of them. I don’t get the feeling that I need to micro-manage my study content– Anki’s algorithm seems to work wonderfully.

With regards to input methods– I don’t know whether mine is rather unusual. I first input the words in Wenlin, (hanzi and pinyin with tone marks (I can’t tolerate numbers for tones)), and then once I am ready, which may not be right away, I copy them over into Anki and begin reviewing. This process of copying and pasting may be slightly slow– but I like it a lot– especially because Wenlin is my primary source for definitions. Do most people just read the definitions out of their class text books?

I’m not sure I fully understand what you mean by ‘cycle elimination.’ Are you saying you want to be presented the same exact card repeatedly (instead of every 10 minutes) until you feel like you can say ‘yes i know it?’ If so, I haven’t tried this– but it seems like It would be impossible to forget a card when no time has elapsed… Yah maybe I don’t understand what you are saying…

Finally, I may very well be naive in saying this– but I have no major complaints about Anki. It has worked really well for me– and I don’t find myself wanting any of the features you describe it as missing. All I want is an iPhone/Touch version.

Look forward to your thoughts/criticisms!

Ironically, you have criticized my being the “final arbiter” in my criticism of Anki’s lack of a crucial feature that my research and experience as a teacher and learner show is essential for the application to appeal to most language learners: study on demand. In a sense, not including that feature made Anki the “final arbiter” on how students should study. Asking users to write a plugin for it is disingenuous. Being able to modify the software doesn’t make it immune from criticism for its existing deficiencies or somehow exempt from review by other people who stack it up against competing software for the same group of learners.

I’m happy if Anki meets your needs, much worse software without any spaced repetition has met the needs of many learners. However, I’m writing reviews for language learners based on what my experience and own research in this topic suggest are a basic feature set that will appeal to the largest number of students. My goal is to spur developers to improve their product and thereby give language learners the benefits of that development.

Oh, you may notice that Anki has, as of this writing, the highest scores I have given out so far in a review. It has a lot of promise!

This entire review smacks me as very conceited. How exactly did you become the final arbiter of what flashcards need or don’t need?

Given that Anki isn’t commercial software, but rather an open project done by Damien in his spare time, I don’t see the lack of any one feature “killing” Anki — if people really want it they’ll go to some other software that allows for it, or — gasp! — write a plugin that allows it in Anki.

Why are students trying to use spaced repetition at all if they’re clamoring for inefficient “cram” review sessions anyway? The whole point of the algorithm is so that you don’t need this sort of “study on demand” feature — the timing is taken care of, at intervals that lead to maximum retention with minimum work.

The design issues are of course difficult due in the multi-platform approach. I sympathize, and personally don’t think they are a big deal, but I’m writing from perspective of a mac user comparing the app to other mac apps. Some of your other points are minor disagreements between us on approach but some I feel strongly about.

You are writing from the perspective of a developer following one approach that is out there in other software. I write as a reviewer from the perspective of someone who has done daily flashcard study for 14 years in three Asian languages, has been surrounded by students doing flashcard study in a dozen different ways, and as someone who wrote an amateur spaced repetition application based on my own method started with regular paper and boxes many years ago.

From my point of view, whatever frills a flashcard app has these are the absolute minimum expectations:

-fast and easy entry (Anki is not bad at all)
-easy editing and organization of entries (Clunky but works)
-graded slideshow method (Excellent)
-study on demand (None)
-interval study (Excellent)
-cycle elimination supported in both study on demand and interval study (None)

Until an app has all those features, it isn’t anywhere near ready for prime time and will fail to satisfy the vast majority of language learners, most of whom are taking language classes and are frequently preparing for quizzes, tests, in-class debates and presentations. Most fly-by-night flashcard applications only tend to these short-term needs. Your more advanced program only tends to their very long-term needs and ignores their real and justified desire to review a group of words when they want. Reviewing any group of words, at any time, is never a bad thing. Complementing such necessary short-term study with interval study, makes it much better.

I’m not going to sit around and wait for my flashcard application to tell me that only now is it ok to review this or that word. If I sit down for my daily study, and interval study only prompts me to study 20 words, then I use the remaining time available to review another 100 or 200 words I’m weak at. Any flashcard app which denies the user this will eventually get tossed by most users or force them, inefficiently, to use more than one app.

Most students have never heard of spaced repetition or interval study. My goal is to give them a flexible environment where they can study in a familiar way and eventually realize that complementing it by a long-term system of memory management also available in the same software can be great for their future maintenance of language skills.

Lack of set management is crippling but your tag approach is consistent with denying students study on demand. Denying them study on demand, however, is fatal, no ifs ands or buts.

Cycle elimination is an absolute MUST - it must only save interval score increments/decrements on the first cycle, then allow them to continue seeing the word as many times as necessary until they get it correct at least once. Then you can do things like delaying 10 minutes or whatever you like. This ain’t rocket science, it is good old fashioned flashcard study and it really works.

I really admire some of the implementation of features in Anki and look forward to seeing its continued dev!

You’re right when you say that Mac users are xenophobes! I guess that anysoftware which works on a Mac will be judged in comparison to other softwareon a Mac, but I think it’s a bit unfair to say “it feels like a(cross-platform) Java application” and attribute that to bad design.Cross-platform development is difficult as each platform has differentusability guidelines, and the Mac platform is considerably different fromLinux and Win32. Making Anki behave like a single-platform application onevery platform would greatly increase the amount of work I’d have to do, andmy time to work on Anki is limited. Also, I don’t even own a Macintosh.

I distribute Anki on the Mac as well because people have requested a Macintoshversion. I know it doesn’t meet all the Mac user interface guidelines (thecross-platform GUI toolkit I use is immature on the Mac which doesn’t help),but I figure that having a version with a (in the eyes of a Mac user) clunkyinterface available is better than having no Mac version available at all.

Some of the other points you bring up:

- I’m not sure if it’s possible to remember the input source in a cross-platform way.

- The edit window used to automatically update in real time, but some people complained about flickering, so I disabled it.

- Cards are deleted when you close the edit window. They’re not deleted immediately to save accidental deletion. Having a ‘del’ shortcut would be nice. File a bug report so I don’t forget

- ‘Study on demand’ is against the principles of a spaced repetition system. The whole idea of spaced repetition is that memories strengthen by allowing them to partly fade before they are exercised. Allowing users to study groups of facts they felt a little uncomfortable with would be counter-productive. Rather than let the user do something which won’t help them, Anki doesn’t make it possible. Perhaps better education should go along side this.

- I don’t agree that there is no ‘cycle elimination’ - that’s exactly how failing cards works. By default there is a 10 or 20 minute delay before the cards show up (again, because waiting actually forces the memories to become stronger), but if you hit the default limit of 20 failed cards, the oldest will be shown again. These numbers are of course all adjustable.

- I don’t see why users should need to be able to tweak individual card intervals. If they have to manually input intervals the system is not working.

- The stats in the edit deck window are smaller due to reasons of screen real-estate. If you have a better idea for the layout, mock it up in an image and post to the forum so people can discuss it.

- You brought up the lack of ’set management’ and say that a lack of study on demand is crippling. I don’t agree. There are basically two models of flashcard programs out there:

stacks/vtrain/etc: let the user do what they wantsupermemo: decide what’s best for the user

Anki errs closer to the latter, because frankly as humans we’re notparticularly good at judging our own memories. It’s hard to suppress thedesire to run over something ‘one more time’ in order to make it stick better,despite the fact that research shows this is pretty ineffective. I’m all forgiving power to the user if I believe it will help them, but I’m reallyunconvinced of the value of “let me choose a group of words to study outsideof the usual spaced repetition system”. The only argument for this is when youare studying under external pressure such as looming exams.

Cheers,Damien