In the “What do you believe but cannot prove” series of questions, one computer scientist gave a good explanation for why he believed there was a better way of programming. He pointed out that the way we program is essentially to take our desired solution, encode it in terms of instructions, and then hope that our encoding (very complicated) matches our intentions.
A better approach should be out there. Computers are starting to get fast enough, particularly with multiple core processors out there, that someone with more time on their hands than I should develop a language in which you tell it what you want at the end, and it’ll just figure out the middle.
Sure it’s a hellish search problem, but us grad students always need good projects to work on, and with BSD Unix and XFree86 essentially done we need some new big task to move forward with.
