Turing’s Machine 223 a) —_— i —_ ir —s \ = Vp ——— ff \ | ae — > i | /f Se ee . tae 4 ai iat js Bs, rr eM a i Targa =| aeeeeelela Bly ae ye, le ME A ee Oe! Mt | Tr, kato a Se ; Py a woey Ge 5 Eeiis Rs Waal | e bia | | g Ben) i lS | as Old Fashioned Relay Mechanism the button to go down. That logic did not exist in the relay banks. If you wanted to improve the logic of the elevator you would need to rip out all the relays and rewire everything from scratch. Turing’s first imaginary machine was set up in the same way. It had a fixed set of hard-wired logic, a rule book. In order to perform different tasks — say addition or multiplication you had to use a different rule book. His revolutionary idea was to write a rule book that told the machine to read a soft-wired set of instructions from the tape and execute those instead. He called this a Universal machine since it could perform any procedure written on the tape. Today we call this software. It is fair to say Turing was not the first to use this idea. Charles Babbage’s analytical engine could read instructions from cards and execute different procedures, but Turing thought through all the ramifications of the idea and made it general purpose, giving us the modern science of computing. It is easy to build a real Turing machine, but by today’s standards it is a little clumsy; a team in Denmark has built one using Lego. You can see a link on my website. Very soon after Turing’s paper was published, a number of people proposed better practical implementations. In 1943, John von Neumann of Princeton University created the architecture for ENIAC, the first stored program computer, developed for the United States Army’s Ballistic HOUSE_OVERSIGHT_015913