[For the entire collection see Zbl 0537.00023.] Optimal binary coding of decimal numbers is studied with respect to minimum cost for hardware and time for the realization of arithmetic operations over n-digit decimal numbers. Using weighted binary coding involving negative weights, $a\sb i=\sum\sp{4}\sb{j=1}a\sb{ij}g\sb j,$ where $a\sb i$ is a decimal representation and the $a\sb{ij}$ are binary representations, there exist 13 self-complementary codes with one negative weight. Further restrictions lead to the investigation of two such codes only, namely 631-1 and 64-21. These codes are analysed and compared with each other and with positively weighted codes, especially with respect to the corresponding logic equations. It is shown that the code 64-21 is practically equivalent to the standard code 8421 with respect to all characteristics.
Reviewer: E.Ihle