At dinner the other night with shunn, bobhowe and our friends Colin, Bill and Christopher, I mentioned the version of Charles Babbage’s Difference Engine #2 built at the Science Museum in London. The story of its construction is told in Doron Swade’s The Difference Engine, which discusses the many challenges that probably would have prevented its construction in Babbage’s time.
In actuality, tolerances and machining methods were not exactly the problem, contrary to what I said the other night. It was more the difficulty of producing parts that were exactly identical to each other. (“Computer-controlled machinery produced the hundreds of repeat parts, the manufacture of which had so handicapped Babbage,” says Swade.)
But beyond that, Babbage’s design was unbuildable — he’d never managed to build it all so was unaware of contradictions and problems in his plans. And once built, there was the entirely different problem of getting it to actually work.”Babbage had not had the experience of fault-finding on a complete machine, and had made no provision for easing the setting-up process or dealing with malfunctions,” Swade says. In a modern age of programming it’s difficult to imagine how in-the-dark Babbage was. “Babbage had made no provision for debugging,” Swade continues. “The whole machine is one ‘hard-wired’ unit,” so if it jammed, which it did frequently, finding the problem was a bear.