esthermations / latinapp Goto Github PK

This is technically LatinWords code but it occurs as an issue in this app so

This is what's currently holding up development here -- using pre-generated words from Wiktionary, we:

1. Don't get all the word forms
2. Use way too much space on storing the data, about 70MB

To get around this, I want to generate word forms on-the-fly as you type. The feature idea is that a noun table or some subset of a verb table would be displayed when you write a word in the table, highlighting your current word so you can see it in context with other conjugations/declensions of the same word.

One issue is how we detect that a word typed is a generated form before we generate them. There are some things I can try here:

• Generate forms once a certain number of letters of a word have been typed. E.g. generate all forms of scrībō once the user has type "scri" or "scr" or "sc" depending on performance.
• Use information from the basic forms of the word to approximate and optimise this.
• Guess word forms based on the suffix of the typed word and try to guess the stem of the word from that.

Etc.

Anyway most of this work is happening at https://github.com/esthermations/LatinWords 'cause it really makes good sense to make this work reusable as a library.

• All proper nouns
• gladiator
• sciens
• clavis

Probably more to be added