I'm not sure then, but hopefully someone here can help you out! Naively 
perhaps it feels like this makes the type inference impossible. Your 
function decomplexify always produces the same output for the same input, 
but what if you had a function that returned a different output (always of 
type DataType) based on a random number generated inside the function? Then 
using that function to annotate field types would be super hairy, right?

Reply via email to