Roadmap
If you have knowledge in language/NLP or would like to help otherwise directly or indirectly with the development of LëtzRead (for example through founding), please get in touch.
Areas for Optimization
Range of Lexical Categories ('parts of speech')
At the moment, LëtzRead only translates nouns. Adverbs & adjective support could be next.
Schedule Activation
Auto switching translations on/off based on a user-setable schedule.
Example: Only active from 17:00 - 20:00.
Open Compound Words
Open compound words are currently not supported. Example 'english teacher' is currently not translated to 'Engleschproff'.
Contextual Meaning / Semantic Understanding
The hardest part of translating any language is semantics. Figuring out this part would require a machine learning ('AI'), specifically NLP approach, and is currently out of scope. These range from words that could be both a verb, adjective or noun ('semantic identification'). Ex: Content classification needed.
Example: The word 'matter' could mean different things depending on context.
- political context: "It's a complicated matter (=Ugeleeënheet) for the government"
- scientific context: "The matter (=Matterie) in the universe is drifting apart"