Null checking can be fine if a failure mode is unambiguous. However, if an operation can fail for many reasons, it can be helpful to carry that information around in an object. For example with URL parsing, it might be nice to be able to know why parsing failed. Was it the lack of protocol? Bad path format? Bad domain format? Bad query format? Bad anchor format? This information could theoretically be passed back using an exception object, but this information is eliminated if null is returned.