On the unique decodability of codes (Corresp.)

A simplification of the Sardinas and Patterson test for unique decodability is given. It is shown that duplicate sequences of code symbols generated by that algorithm can be omitted, resulting in a more efficient algorithm with respect to computer storage and time required to carry out the test. This simplicity is employed in deriving an improved upper bound on the number of sequences of code symbols that must be generated in order to determine whether or not a given code is uniquely decodable. This easily computed bound serves as a measure of the amount of computer storage and time required to carry out the test for any given code.