GENERALIZED CODING AND UNIFORM FINITE MEMORY CODES
CALIFORNIA UNIV BERKELEY ELECTRONICS RESEARCH LAB
Pagination or Media Count:
A general framework for encoding and decoding is presented which includes block coding. The key concept used with the generalized codes is that of decoding rate. A weak converse is proven using decoding rate which shows that channel capacity for the generalized codes is the same as the usual block coding channel capacity C for a discrete memoryless channel. Uniform finite memory codes are defined from the general frame work after several motivating definitions of properties which seem natural to require of any code. Channel capacity C sub u is defined for these codes but what its value is remains an open qu. A class of channels is given for which C sub u is nonzero for each member of the class. From the converse it is known that C sub u is equal to or less than C.