Aurorum / latin-vocabulary-tester

Provides a way to test Latin vocabulary, noun declensions, and verb conjugations based on the A-level and GCSE syllabus, while containing various Latin word games, including Scrabble and Wordle.
https://latinvocabularytester.com
8 stars 2 forks source link

Redundant data in vocab lists #20

Open rayzchen opened 2 months ago

rayzchen commented 2 months ago

Under vocab-lists/ some word entries in the JSON files have explicit declined or inflected forms of the word, while others don't. This leads me to think that perhaps it is not necessary for many of the regular verbs. Do the explicit inflected forms serve any purpose or can they be autogenerated from the other principal parts?

It also seems that some irregular verbs don't have these explicit inflected forms.

Aurorum commented 2 months ago

Hi @rayzchen, thanks for your question!

Do the explicit inflected forms serve any purpose or can they be autogenerated from the other principal parts?

If an item doesn’t contain any inflected forms, then that word won’t appear in the declension/conjugation tester, nor will the vocabulary tester ask for an inflected form. As such, the irregular verbs are generally excluded, not least because the code currently relies on all the inflections being filled, and some verbs can’t be conjugated in certain ways.

It’s a very interesting point as to whether inflections can be automatically generated from the principal parts rather than hardcoded for each item. To be honest, I suspect that they can, but there’s probably all sorts of irregularities to watch out for - deponents, irregulars, plural nouns etc.

The main advantage with the hardcoded inflections is that I (and the user!) can be certain that they’re correct. That being said, it would certainly reduce the load bundle to autogenerate them, so I’d love to have a think about how to achieve that without risking any incorrect answers.

Thanks for raising this interesting issue, and let me know what you think! :)