The char datatype can be treated as its unicode equivalent without typecasting or converting. So it's is evaluating the chars based on their unicode equivalents. If you're not familiar with Unicode you definetly need to research it.