Masatoshi Shima (嶋 正利 Shima Masatoshi , born August 22, 1943) is a Japanese electronics engineer, who was one of the designers of the world's first microprocessor, the Intel 4004, along with Federico Faggin, Ted Hoff, and Stanley Mazor.
He studied organic chemistry at Tohoku University in Sendai, Japan. With poor prospects for employment in the field of chemistry, he went to work for Busicom, a business calculator manufacturer. There, he learned about software and digital logic design. When Busicom decided to use LSI circuits in their calculator products, they approached the American companies Mostek and Intel for manufacturing help. The job was given to Intel, who back then was more of a memory company and had facilities to manufacture the high density silicon gate MOS chip Busicom required. Following Marcian "Ted" Hoff's initial conception, formulated in 1969, Shima later helped design the 4004 processor, working at the Intel offices for six months- from April until October 1970- with Federico Faggin, the project leader. His company then sold the rights to use the 4004 to Intel, with the exception of use in business calculators.
Intel then designed the 8008 (architecture by Computer Terminal Corporation, design by Federico Faggin and Hal Feeney). Shima was then employed to implement the transistor-level logic of Intel's next microprocessor, which became the Intel 8080 (conception and architecture by Federico Faggin), he was not involved in creation of 8088 and 8086. Shima moved to Zilog in 1975 and, using only a small number of assistants,[1] developed the transistor-level and physical implementation of the Z80, under the supervision of Faggin, who conceived and designed the Z80 architecture to be instruction set compatible with the Intel 8080. This was followed by the same task for the 16-bit Z8000.[2]
According to co-workers from Intel, Faggin's method that Shima used was to design all logic at the transistor level directly and manually (not at the gate and/or register level). The schematics were therefore hard to read, but as transistors were drawn in such a way that they suggested a "floorplan" of the chip,[3] it actually helped when making the physical chip layout. However, according to Shima himself, the logic was first tested on breadboards using TTL chips, before being manually translated into MOS transistor equivalents.