This isn't ideal, because it forces each successive function that takes a Result type to add handling code in the case of an error.

fn(fn(fn()?)?)? is a bit gnarly but better than duplicate code like that, imo.