The Bet Behind Formula Translation
Fortran, short for Formula Translation, is usually filed away as the old language, the one scientists used before something more modern came along. That framing misses what the language actually settled. When IBM released it in 1957, the team led by John Backus was not just giving programmers a friendlier notation for writing formulas instead of raw machine instructions. It was staking a specific, testable claim: that a program written in something closer to ordinary mathematical notation could be compiled down to machine code running about as fast as code a human had hand tuned instruction by instruction.
At the time, many programmers doubted that claim outright. Hand written assembly was assumed to be the only way to get real performance out of a machine, and a compiler was expected to waste cycles a careful human would not. Fortran early success argued the other way, and it argued it in the language that the people who most needed the speed, scientists and engineers running numerical calculations, actually used and trusted. Every high level language written since has inherited the ground Fortran won in that argument, that a programmer does not have to choose between writing in a human readable language and getting machine level speed. Fortran itself is still active today in high performance scientific computing, still doing the specific job it was built to prove could be done at all.