It is difficult to identify the subject of this paper, which claims to be a study of cellular processors (defined as arrays of “microprogrammed Boolean processors connected according to the von Neumann neighborhood”). Katona claims that three levels of languages are necessary in order to program such an array: “a microassembly language, cellular assembly language, and a cellular graph language.” The application shown is the execution of a multiplication of Boolean matrices on a cellular array, which closely resembles a systolic execution of this multiplication. The use of the three languages is not convincing.

The paper touches on diversified problems--systolic computation, I/O problems for matrix computation, and use of a microprogramming level--but nothing is really clear. The conclusion and the bibliography are nonexistent. Overall, I do not consider this paper well focused or technically sound.