digitapex / DerDieDas

Android app for learning genders of German nouns
MIT License
14 stars 10 forks source link

some words can have several genders: support? #2

Open guillaume-d opened 5 years ago

guillaume-d commented 5 years ago

Some (a few?) words in German can have different genders depending on the meaning, like der/das Teil, for which DerDieDas the moment only accept das (although AFAIK der might be the more commonly used). AFAICU you data source for genders is Wiktionary, which gets it right: https://de.wiktionary.org/wiki/Teil The trickiest example I know is der/die/das(!) Band: https://de.wiktionary.org/wiki/Band For that one there is some disambiguating text in the Wiktionary article after the gender, which could be added to have different questions, like: "Band (Musikgruppe)" -> die, "Band (Reifen)" -> das, "Band (Stoffstreifen/Plastikstreifen)" -> das, etc.

As handling this correctly in a generic way seems like a lot of work, my suggestion is:

  1. treat Teil, Band and the other (few?) cases manually in the DerDieDas app itself by either replacing the single gender with more variants as outlined above, or at the very minimum delete multi-gender words altogether once spotted, as keeping only one gender would be teaching something wrong / misleading
  2. maybe extend https://github.com/digitapex/ParseWiktionary so that it can at least detect multi-gender words to get a feeling about how much words may have multiple genders

Also:

guillaume-d commented 5 years ago

DerDieDas only accept das für Band, so I guess only the 1st entry from Wiktionary is taken, right?