Signed number representations, two's complement and overflow, binary add/subtract, and decoding or encoding a 32-bit IEEE 754 float, all checked with a program.