Comparison of iPhone Edict software

General Navigation

The Japanese app has a main scrolling menu, and features are available from here. This organization is very logical and extensible, and very friendly for new users. The drawback is that expert users need a lot more taps to move between the most commonly used screens (from lookup to history, for example) than in Midori or Kotoba!. Midori and Kotoba jump right in the middle of things, the main screen is the search screen, and everything else can be reached from there. This is a more efficient, but less friendly approach.



Copy-pasting and jumping around is very poorly supported in Kotoba in general. You can usually only copy entire lines, and hardly anything offers hyperlinks to related content. There are also very few “related” lists to be found anywhere.


Midori displays every screen like a webpage, so you are free to select and copy whatever you want. Meaningful items are also usually (but not always) hyperlinked, so navigation is quite comfortable. Related words and kanji compounds are offered and conveniently navigable.


Japanese doesn’t allow the free copying Midori does, but it is very well hyperlinked, so you hardly ever miss it. In addition to the kanji compounds, Japanese also lists word compounds.

Information display



Dictionary searches


Result Display

Results Structure

Result Ordering


Midori follows the model of “real” Japanese dictionaries, where the entries are the hiragana words, and the entries list the possible kanji transcriptions for that word.

The result list contains the hiragana as well as the kanji.

Kanjis are not directly searchable in Midori, they are linked from the word result pages. This requires an extra step to look at the kanji (if that’s what you need), but it reduces result clutter and helps avoid confusion.


Japanese’s model is pretty confusing. The entries for single-kanji words are somehow merged together with the kanjis themselves, so you can never be sure what the real reading of the word is. The entries are defined by a kanji transcript, and other potential kanji transcriptions are listed as “alternatives”. This has some interesting (undesirable) side effects in the result list, which sometimes lacks hiragana, and sometimes has unexpected hiragana for a kanji word.


Midori requires you to explicitly select the direction (E->J/J->E) or the names database. This helps searches go faster, and reduces the clutter in the results, but a bit cumbersome if you often have to switch directions. The other applications look in all three dictionaries.

Character entry



Kotoba! wins hands town for non-handwriti

Word lists


In addition to the “usual” kanji lists, Japanese offers various word lists grouped by JLPT level, lexical categories (phrases, nouns, counters, etc.) and common topics (agriculture, law, sport, etc.).

The problem with the JLPT lists is that it’s unclear which reading(s) are required to be known, and which are rare/advanced. To make matters worse, in the word list the hiragana readings seem somewhat random (another side-effect of trying to make dictionary entries based on kanji instead of hiragana). The other fundamental problem is also apparent: the “word” list excludes single-kanji words, and the “kanji” list shows arbitrary readings.


Midori only has Kanji (by frequency, by JLPT level, by school grade), and hiragana/katakana lists. Like Japanese, Midori does not indicate which kanji readings are JLPT and which are not. Word lists are completely absent.

The Kanji list looks best of all three, as the characters are big and all the readings, as well as the meaning are available at a glance.


Kotoba mixes the kanji lists with the kanji lookup, for a spectacular result (see above). Word lists are not included.

See Also