From #32

For the order what I had intended for validation is to have two modes of parsing, strict and loose. This was intended to allow reading corrupt files or files from a less compliant encoder. It's a secondary/tertiary goal of the project, I think lets focus on correctness first.