Fool’s Flashcard Review

Kanji Flip is a powerful flashcard application targeted for students of Japanese Kanji. Using the graded slideshow approach offers an excellent interval study implementation that will serve even the worst of memories well. The developer, Andre Khromov, also offers a range of other applications based on the same basic features, including applications focused on Japanese vocabulary, the Russian language, and the Korean writing system.

While the feature range is narrowly restricted to serve its primary task, this enables Kanji Flip to do the one thing it sets out to do admirably: help you memorize over two thousand Japanese characters. It is somewhat more expensive than other offerings, could benefit from some minor improvements, but given its powerful interval study feature, as of this review Kanji Flip is currently the clear leader in iPhone/iPod offerings for students of Japanese characters.

Application Name: Kanji Flip
iTunes Application Store Link: Kanji Flip
Version Reviewed: 1.2
Software License: Commercial (about $6)
Review Date: 2009.01.26
OS Tested: iPod Touch 2.2

Note: This review is primarily from the perspective of language learners. See the Terms page for an explanation of the technical terms used in these reviews. See the Basics page for a list of basic features found in flashcard applications useful to language learners.

Kanji Flip is justly confident in its ability to get you going with kanji characters. Even the icon (which, given its blue on red color is probably the ugliest my iPod touch currently sports) has chosen the character 叡 or えい (C: ruì S:睿), which can mean shrewd, profound, or astute, to declare the wisdom of its approach.

At launch Kanji Flip offers a standard list of sets of characters to study based on those in each of four JLPT levels ranging from 80 characters for level four to 2230 characters for the highest level one. You may also practice the hiragana and katakana syllabaries via a special kana practice zone. Finally, if you wish to add your own kanji to the list, in the unlikely event that the most commonly used 2230 characters are not enough, you can also add them. There are not, however, any options for set creation or set organization. Given the highly targeted nature of the application, however, this is not essential.

Kanji Flip is, from start to finish, an interval study powered application. Because some users may not be familiar with this style of study it is wonderful that it includes built in help that explains how the process works. However, the exclusive focus on interval study means it offers absolutely no study on demand so students cramming for tests or studying along in a textbook may with to consider other alternatives. You can, however, browse through characters of one of the offered lists if you enable the “browse mode.” Although a very minor issue, the browse mode button is annoyingly “sticky” and requires a swipe of the finger rather than a press, unlike most other switch buttons in the iPhone/iPod interface.

Kanji Flip is built for long term study and there is, in fact (and the developers website acknowledges this) no real reason a self-study student should begin studying one of the lower levels unless one is targeting a particular JLPT examination.

Flashcard Interface

Flashcard study is done in the following way: A kanji is shown and the user taps the “show answer” button or the large white space on the card. Unlike the fully touchable surface of Lima Sky’s “Kanji” application and its “known” button located in the easier to reach upper half of the screen, these buttons in Kanji Flip are all located in the harder to reach bottom of the screen, which is a further stretch for the thumb when holding one’s iphone/ipod. This is hardly noticeable at first, but if one goes through a hundred or more reviews each day, one begins to feel the pain in one’s hands. I hope Kanji Flip will consider making the whole surface touchable to show the answer, and locate the buttons in the easier to reach upper half of the screen. Once you see the answer you can indicate, in true graded slideshow tradition whether you have recognized the character. Unlike some less well designed flashcard applications, Kanji Flip immediately moves to the next card when you have marked it correct or incorrect.

Interval Study

Using an interval study algorithm which is, unfortunately, not made transparent, the Kanji Flip application slowly introduces new characters as you master a small set currently in the cycle. As your memory of the characters improves, the characters very slowly begin moving down through the “Recent”, “Older” and “Ancient” interval stages. This process will seem extremely slow to the user at the start, but I think Khromov has made an excellent algorithm which errs on the side of the conservative: he is pessimistic about our memories and my years of language study show this to be an 叡 approach.

Khromov has also designed Kanji Flip in such a way that you are never finished. You will never “know” all the kanji and be able to call your task complete. Kanji Flip will forever drag out the oldest characters for you to review. This is an absolute key and basic principle of long term language study that is sorely missed by even the biggest commercial software developers who are ignorant of this basic fact. As the developer wisely puts it in his website FAQ:

“Please realize that the goal of learning (and this program) is to continually strengthen your mind. There is no such thing as “finished”. The kanji in a set will never all pile up in the Ancient area. After you exhaust all the untested kanji, the oldest previously tested kanji will be put back into testing rotation – even if you answer it correctly. This is done on purpose.”

I do think, however, that the algorithm can benefit from some kind of date tracking. While the application should never conclude that the user is done forever with the task of learning characters, it is not necessary to drag out characters before they are likely to be forgotten. For example: If I have a word which finally reaches the “ancient” category and has been reviewed over two dozen times (say), then, the following day, if there are no other words “up” for review, it isn’t necessary to force me to review that card which has just recently reached the “ancient” category. I might use that time to practice other cards or other languages in another flashcard program, for example. Since Kanji Flip is not a general use flashcard application it cannot assume users are only learning its material in the way other general flashcard programs, where you can practice a large variety of information, may be more justified in assuming. A word in the “ancient” category can probably go several weeks without being forgotten – weeks during which my time is better dedicated to learning new information, say working with Khromov’s Japanese or Russian vocabulary applications instead. Thus, while the developer is completely correct in maintaining that we never permanently know any words, it is also not the case that, at any given time, we necessary have forgotten any individual character.

A solid interval study implementation keeps in mind the time-to-forget (TTF) for any given unit of knowledge and, ideally, does not prompt you to review a word until it has nearly reached the threshold of that TTF, as calculated by whatever algorithm chosen by the developer (and ideally tweaked with some input from the user based on their own experience and strength of memory). This allows students to make maximum efficient use of their study time, especially when a large percentage of target knowledge is at a distant TTF and they wish to focus on new material elsewhere.

It is not clear to me the exact interval study method being implemented, but my testing so far shows that every half dozen or so correctly marked characters will result in one new character being introduced. Poor performance is brutally punished with a high repetition rate, and that is as it should be. Given the fact the program clearly is doing a lot of tracking, I would very much like to see greater transparency in future versions by offering the user a way to inspect certain statistics of study that are clearly being tracked (e.g. total reviews for a card, total marks correct, incorrect, and a more precise indication of card progress). The colorful “progress” table at the right is useful but large and sometimes distracting. I would recommend the following improvement for future reviews: Allow this area to be blank, so there is no distraction, but allow a tap to toggle between the progress table, and specific data on the currently viewed card, and then back to blank. Otherwise, I find myself too often looking over at the progress table instead of concentrating on the card at hand.

Supplemental Vocabulary

There is vocabulary attached to each character, which is excellent. There are two possible improvements that might be considered for future development: 1) Using a method similar to Lima Sky’s “Kanji” application show the example vocabulary before the answer has been shown, but hide their pronunciation and definition, only to be shown when prompted for the answer. That way a user can easily practice vocab along with the kanji. 2) Many Japanese kanji have verbs or adjectives based on the kun-yomi reading of the character. While the kun-yomi readings are shown for each character, too often the definitions for these kun-yomi words (sometimes quite obscure) are not included among the example words. Many kanji textbooks use a method I recommend: among the example words, include the most frequently used kun-yomi word.

Other Comments

In the realm of statistics, Kanji Flip is great in that it shows you not only the current number of words in each of its general interval stage categories, but also the percentage of words answered correctly in any given study session. However, as noted above, it would be nice in future updates to have a little more of an overview of individual card statistics or overall study statistics in some location.

While it is understandable that synch or update/export options are not available in this highly target application, this does leave the user feeling very vulnerable. Learning or reviewing over two thousand characters can take several months or more than a year of almost daily study. The possibility of data loss or other problems with one’s iPod/iPhone are real, and I think all strong data-intensive (in this case, the data being about one’s study in the form of interval statistics) applications should offer some way to back up the data (beyond the standard iTunes backup of the entire device) either by wi-fi synch or email. Given the highly professional nature of this excellent application and its potential for intensive long-term use, I hope that the developer will consider offering such a capability in future.

My comments above should largely be seen as recommendations to make an already strong application stronger, and I believe much of what is said here is equally applicable to some of Khromov’s other flashcard application offerings. Kanji Flip is way ahead of its competition for kanji study software. I hope the developer continues to strengthen its features in order to maintain its solid lead.

Import: None.
Export: None.
Non-Roman Scripts: No problem
Modes of Study: Graded Slideshow
Media and Frills: None.

Entry Creation: Can add kanji to the
Entry Editing: None
Set Organization: None
Flashcard Study: Excellent
Interval Study: Excellent 4 level system, very conservative, but no transparency at the level of individual characters
Formatting: No customization possible.
Design and Feel: Good, progress window a bit distracting, icon is nasty looking.
Statistics: Good, session average shown and overall interval progress indicated.

Golden Coxcombs: 9/10 Currently the top of its class.