Jeffrey's Fuzzy Kana Search

What is important to one language's ear might sound insignificant to another's, and it works both ways. For this reason, I have built in the idea of fuzzy searching, whereby the search will often be able to ignore those linguistic characteristics of a word's spelling (in romaji) which might be missed by an English speaker.

For example, the capital city of Japan is not ``tokyo'', but ``toukyou''. The long vowels are very important in Japanese, yet the English speaker is used to glossing over such details as not being important to our speech.

To help the non-native speaker find words more quickly, I've implemented what I call a fuzzy search. A fuzzy search will attempt to mitigate the difficulties non-native speakers have identifying the exact Japanese.

In a fuzzy search:

• the length of vowels (and if long, how they're lengthened) is ignored -- For example, ``to'' is considered the same as ``tou'', ``too'', or ``to-''.
• the presence or absence of any small tsu is ignored.
• the dual characters with the ji sound ( ) are considered to be equal.
• the dual characters with the zu sound ( ) are considered to be equal.

As an example, consider the fuzzy search of ``tokyo''. Among other, the following will be returned:

tokkyo

special permission; patent

tokkyou

convex lens

toukyou

Toukyou (pl); Tokyo; current capital of Japan

There is a word in Japanese, , for someone whose kanji-writing ability has suffered due to over-reliance on the kana->kanji conversion systems used to input Japanese text on a computer. They merely need to recognize the correct kanji, so their ability to actually produce it diminishes. Don't let this kind of thing happen with your pronunciation

One note: a fuzzy search does not include certain grammatical points of aural confusion:

• ``wa'' will not match (use ``ha'' to match it)
• ``o'' will not match (use ``wo'' to match it)
• ``e'' will not match (use ``he'' to match it)