When Backus and a small band of IBM colleagues began their quest in 1954 for a programing system that would enable a computer to produce its own machine language programs, they weren't always sure what they would come up with. In 1967, Backus remembered: "As we began to solve one problem, it split up into others we hadn't foreseen. In January 1955, we said we would have it in less than a year. Finally, we did it in 1957.