Larger corpus

[madig]

It would be interesting to consolidate various language data repos into this one maybe, so we can use it in e.g. regression testing, as a source of words and sentences to test-shape for a font. Real-world data to use for testing.

Data sources could be the Android Open Source Project UI strings, all of Wikipedia, the Universal Declaration of Human Rights, etc.

[simoncozens]

Superseded by #48.