Comment by InitialLastName

One would think it wouldn't be too difficult to run spell-check + commonly-misinterpreted characters fuzzer on OCR material.