v0.17.0 Release

Today I updated Elephant to 0.17.0.

Let me start with an apology: The rollout wasn’t announced and took much longer than expected. Both of those will improve next time with an upgraded DB instance and a rollout announcement.

If I had to give this a title, it is the

Big Vocabulary Quality Update

This version changes a lot of the internal structure; here’s a short overview.

Idioms and Proverbs

Idioms and proverbs are detected differently and shown in the sentence analysis on their own.

Better Phrase and Word Matching

There are a lot of analyzer improvements in this release that should create more reliable word and phrase detection. This should improve the overall coverage of sentences.

Classification

With this release the preview from 0.16.1 of classification based on the National Institute of Korean Language (NIKL) standard curriculum recommendations is fully supported. ~99.5% of the NIKL vocabulary is recognized - the dictionary obviously contains more than that - and all recommended grammar has basic support in Elephant.

Practice improvements

Elephant should now group meanings that are so close together that it doesn’t make sense to study them separately. The largest interaction is, however, with the previously mentioned classification system: Instead of having to track everything yourself, you can pick your study level and Elephant will learn your vocabulary as you exercise. Vocabulary you correctly use in other exercises will be added to your study deck and can be trained and tracked explicitly.

This should simplify starting with exercises a lot!

There were also a number of smaller improvements to practice navigation, listening, with first work on larger text support: dialogs and passages of text.

The next release, 0.18, will focus on improvements to grammar.

David