I like what the error handling achieves, it is actually readable way to understand the divergent control flow paths (and probably majority of code is read more than written), but I do not enjoy writing the initial boilerplate, so that's good to hear.