Faster to parse as a developer. I can't speak to whether computationally it is faster.

Additionally, doesn't this somewhat also bring in the methodology of dealing with the error cases first?