Home

Welcome to the Dark-Gate wiki!

The name comes from http://en.wikipedia.org/wiki/Dark_Gate

The core purpose of Dark-Gate is to facilitate sentence mining in Anki. The focus is for the Japanese Language, because that is what I am studying. Ideally, Dark-Gate will be able to find the best sentence for you to study next. Based on what you know now (mature cards), and what you are studying now (young cards). What is best should be based on:

1. Vocabulary
2. Grammar
3. Size of the sentence

To be able to do this, Dark-Gate needs to be able to recognise words in a sentence, detect possible grammar points/patters and make judgements on sentence size.

Vocabulary

Potential problems with different meanings of the words, but I guess this will be minor. Look at what the user is currently studying and try to match words that the user is already studying. When introducing new words, consider:

1. the frequency of the word
2. when it is traditionally introduced in school, and
3. the kanji compounds (more on this later)

Sentences

This is a little unknown at the moment. Ideally, there would be a big list of grammar points in the form of regular expressions, and everything would be categorised into one grammar point or another through parsing the sentences. But maybe we just have to do a fuzzy matching heuristic. Here, we would just have to observe the current sentences and try and get similarities. For example, this sentence has noun, partial, verb. etc and we would match like this. When showing a new sentence, we could say, well, we don't know what type of grammar point, but it is similar to these other sentences.

Size of the sentence

This is easy, smaller is better.