Researchers discover computer pioneer Konrad Zuse's long-forgotten Z9, the world's first program-controlled binary relay calculator using floating-point arithmetic.
The following letter was published in the Letters to the Editor in the October 2012 CACM (http://cacm.acm.org/magazines/2012/10/155547).
In his news story "Lost and Found" (July 2012), Paul Hyman characterized Konrad Zuse's Z9 as "the world's first program-controlled binary relay calculator using floating-point arithmetic." This description is not correct but should indeed be the other way round; the Z9/M9 was the only one of Zuse's computers to use binary-coded-decimal fixed-point arithmetic.
Zuse used binary floating point from the time of his earliest computer designs, because his own thorough analysis showed binary representation reduced the complexity of the arithmetic unit and that floating point is adequate for engineering calculations, which, as a civil engineer, is what he primarily had in mind.
Among the pioneers of early computing, from Babbage to Aiken to Wilkes, Zuse alone used floating-point arithmetic; his general-purpose computers Z1 (1938), Z3 (1941), Z4 (1945), Z5 (1953), and Z11 (1955) all used binary floating-point arithmetic. Beginning with the Z22 (1958), the computers developed by the Zuse Company used binary fixed-point arithmetic, implementing floating-point arithmetic through microprograms.
Zuse invented a format for binary floating-point numbers similar to that of IEEE 754, using it in his very first machine, the Z1; Donald Knuth attributes the invention of normalized floating-point numbers to Zuse. The Z3 included representations for 0 (zero) and (infinity). Operations involving these "exceptional" numbers were performed as usual, as in 0 0 = 0 and + 5 = . Operations with an undefined result (such as 0/0, , and /) were detected automatically, signaled by a special light before the machine stopped.
Zuse discussed the question of binary vs. decimal arithmetic with Howard Aiken, writing, "We also had some differences of opinion on technical matters. Aiken advocated the decimal system and had developed very beautiful codes for decimal numbers using several bits. I was more a proponent of pure binary representationin any case, at least where large scientific computers were concerned. However, I had also used encoded decimal numbers in the mechanical solution for the punch card machine."(1) The "punch card machine" was the Z9/M9.
Jrgen F.H. Winkler
(1). Zuse, K. The ComputerMy Life. Springer, New York, 1993.
Displaying 1 comment