The text is probably the output from an OCR program that was run on a printed copy of the test, and the user didn't scan the output for the inevitable errors that these programs produce.