Nihongo Pera Pera (Fluent Japanese)

These were originally intended to be posted as comments on the review pages. However, I had mistakenly thought I had disabled comments on the review pages when in fact I had not; hence, I (Kaeru) am re-posting the comments here. (I intended to have no comments on the review pages because when I update the review, the comments don't fit the page any more, and I don't want people to have to read through a series of comments and sort through what is still relevant and what isn't.)

I have updated the review in some areas in response to these comments.

http://www.nihongoperapera.com/spaced-repetition-software-platforms.html

I don't understand the comment 'the web is too slow for normal use'. Up to ten cards are cached in AJAX and you can use keyboard shortcuts to review. The experience should be largely like the desktop version.

As for data rates in Japan, the cell phone interfaces are optimized to use as little data as possible. Moderate review is quite inexpensive. And if you have a long daily commute, the ability to review on your cell phone is a great time saver. Furthermore, most cellphone contracts allow you unlimited packets at a capped rate of 4000yen (~$40)/month. Saying "I would need a raise" makes it sound like the feature is essentially useless, which I think is unfair.

http://www.nihongoperapera.com/spaced-repetition-interface-comparisons.h...

(Hi, I'm the Anki author)

You can display the 'category' in Anki too by including the tags in the question format.

You can also display a horizontal divider between the question and answer by adding it to the question or answer format.

The next version of Anki (0.9.6) has added the ability to hide the last question and answer if you wish. Many people appreciate that functionality, however, so it remains on by default.

As for your complaint about the vertical buttons, that allows explanations for people who are not familiar with the SRS algorithm. If you prefer a more compact format, it takes about 2 seconds to change it. You make it sound like a chore. :-)

http://www.nihongoperapera.com/spacing-algorithm-reviews.html

(Hi, I'm the Anki author)

I disagree with your assertion that displaying the next scheduled time is a bad thing. If you trust the program to schedule things properly, your answers shouldn't be related to the predicted next times. However, I understand that some people may prefer not to know. The next version of Anki (0.9.6) allows you to hide the predicted scheduling information if you wish.

As for failing cards, displaying them again in 10-20 minutes helps cards transition from short term memory to long term memory. I think you'll find that Mnemosyne behaves in a similar way, by showing you only a limited number of new cards at a time and looping through them until you've been able to adequately remember them. Mnemosyne's 'max number of cards to show at one time' option is effectively the same as time-based limit. Furthermore, a 'final review' as you describe is not particularly encouraged by the SuperMemo folks, and they actually say that as it's not important it should be disabled if you're short on time.

You complain about Anki's spaced repetition system being too configurable. It's possible to change the initial intervals and the amount of spacing to apply to similar cards, but the core of the algorithm always behaves the same. The options available to the user make it hard for the user to shoot themselves in the foot.

As for card scheduling being per-second rather than daily, there are people who appreciate the trickle of cards, and say they much prefer cards appearing in small numbers during the day to having a large number of cards due once a day. Again, this comes down to personal preference.

As for the research element, Anki is not targetted specifically at Japanese learners. It includes some support for the Japanese language as I have experience in that area, but there is also support for Chinese reading generation, and Anki can be used to study basically anything. The algorithms used in Anki are based on the existing research available, just as the ones in Mnemosyne are. Mnemosyne collects some data on the user's reviews, but all that data is being recorded by Anki too. I think it's a bit unfair of you to say Mnemosyne 'benefits immensely' from the 'research' - Mnemosyne is largely based on the SM2 algorithm, just as Anki is, and all the data mined by the Mnemosyne project so far has not resulted in anything so far. Furthermore, the same data is recorded in Anki decks and is available for future data mining if the need arises.

As for 'credit where credit is due', I should have said 'pretty functional spaced repetition implementation' on that page. That was not an attempt at deception, just a poor choice of words. I've updated the page.

http://www.nihongoperapera.com/stability-documentation-support-reviews.h...

(Hi, I'm the Anki author)

There are 5 screencasts which introduce most of the core functionality of Anki. These function as the base documentation. I'm not sure if they existed when you wrote your review.

http://www.nihongoperapera.com/spaced-repetition-review-conclusions.html

(Hi, I'm the Anki author)

You mentioned earlier in the article a problem with Anki showing questions when it shouldn't. That was a bug which has been fixed, but it was fair to say that such a bug probably shouldn't have made it into an official release. However, as Anki nears 1.0 it has become quite stable. I think it's worth pointing out that you reviewed Anki during a transitionary period, and the problems related to that have been ironed out for some time. So I'm not sure if it's fair to describe it as 'a little buggy' anymore.

As for the 'interface problems', as far as I can gather from the text of your article, these seem to be stylistic preferences related to a separation of question and answer, and not displaying the last card or next scheduled time. As mentioned in previous comments, all of this is configurable.