Data Exchange should determine if a file being loaded is in UTF-8 or MARC-8 format before loading. If it is MARC-8, it should be converted to UTF-8 before loading.

There are still a lot of MARC-8 records out there, and it is not easy for libraries to determine which format any given record file is. When MARC-8 records are loaded diacritics and special characters come out completely messed up, making those terms unsearchable and unreadable. Our Sierra server has a lot of these records, and they are still being added because libraries can't get vendors to provide UTF-8, and they don't know how to fix them before loading.