Vortmaro

It's been a while between posts, but that doesn't mean that nothing has been happening.

Work to support new languages is underway. For example, #Greek

The system can now parse separable verbs and novel compound words from German text. There's still testing and tweaking to do, but beta testing will hopefully open soon.

@languagelovers@a.gup.pe what languages are you learning, and what tools do you use to help with #reading in them? #LangToot #LanguageLearning

Current tasks:

* Support registrations for beta testing
* Add a simple tool for learning (#Japanese) #hiragana and #katakana syllabaries
* Start work collating (free, online) #German reading resources (recommendations welcome!)

Wiktionary has an excellent range of word definitions for German, but not all words can be found in the dictionary. For this reason, Vortmaro now also uses Wikipedia as a data source. Here are some examples from DW articles.

Vortmaro is designed to give grammatical information relevant to both beginners and ongoing learners, with each word entry specifying the base word and relevant attributes (such as case, gender, plural).

For example, for the German word Kindern, Vortmaro indicates that it comes from the base word Kind, and is in the dative case and is plural. This information is extracted from Wiktionary, which has excellent declension tables.

Hello world
Saluton mondo
Γειά σου Κόσμε
こんにちは、世界

Vortmaro now handles Wiktionary entries with multiple genders listed for the same word (e.g. where the word's gender is a regionalism or speakers use multiple genders interchangeably).

At this point German language support is (at least theoretically) feature complete, which means the main focus now will be on preparing the system for user registrations for beta testing.

Update on Wikipedia as a data source: Vortmaro now provides images and supports disambiguation (where an entry in Wikipedia lists possible alternatives which could be meant by a single word or phrase).

Toying with the idea of just turning on support for all* the languages and declaring them to be in alpha. 😅

Once flashcards are working and the Firefox addon is available again, of course.

*Really only those which are written left-to-right and include spaces between words, and that the Wiktionary parser can handle without significant work.

Much of the work done to support German will be useful for supporting other languages, but each language presents its own challenges. So in addition to the remaining work to support German, exploratory testing will begin on other languages, starting with French and Spanish.

Recent improvement:
Include usage tips found for certain words in Wiktionary, e.g. for the German preposition "trotz": "+ genitive or dative"

Some of the things left to do before opening up German beta testing:
* Try to define proper nouns using Wikipedia when there's no Wiktionary data.
* Give partial results instead of none, e.g. for the made up word "Wonkyflonk-itis", give a result for "-itis" even when "Wonkyflonk" can't be found

This week there has been some experimentation with other languages (noting that Vortmaro development is mostly focused on German at the moment).

It seems that a reasonably straightforward architectural change (adding DB support for the layout of declension tables for each language's word roles) should make it simple to add proper support for new languages - that is, with the level of grammatical detail which is a core part of Vortmaro.