Hats off to the father of FORTRAN
The headline reads John W. Backus, 82, Fortran developer, dies. May he rest in peace.
FORTRAN, short for "formula translator", was developed in 1957 and fifty years later, it is still of immeasurable value to researchers at universities all over the world.
Today, we think of Education IT as that group within a school or university which provides workstations and network infrastructure for student use. In truth, Education IT -- especially at the university level -- is a great deal more than that.
The very first university computers were research computers -- often under the control of a small handful of 'experts' in a single discipline with a very specific need. For instance, our computer center is named after an astronomer who pioneered computing on our campus. He probably used the computer for calculating the orbits of asteroids for the Minor Planets Center. There are still a handful of chemists on campus that remember those days and have their own 'war stories'.
I've heard some of those stories -- about the IBM 650 in the basement of the Library with 60 words of core memory, 10 of which was reserved for the OS (such as it was). Hollerith punch cards, EDCIDIC codes, and drum storage, with 50 words per track was all you had -- and, of course, cabinets full of vacuum tubes and blinking lights.
FORTRAN represented the first shift away from the 'computer scientist' and toward the computer user -- who, in those days was a scientist in his own field of expertise. No longer was computing limited to people who could make sense out of strings of ones and zeros -- or even those who could read assembler and keep track of registers loaded and operated on using post-fix notation (also known as 'Reverse Polish Notation' -- after the Polish mathematician who invented it).
FORTRAN made it possible for most anyone with an understanding of mathematics to construct programs to address most any problem that can be represented mathematically. As in its day, today FORTRAN is used for any re-iterative mathematical process which converges on a solution over an extended period of time.
Despite its age, FORTRAN continues to be extremely valuable for these kinds of problems -- in large part because of its simple yet powerful instruction set with which virtually any data structure can be constructed.
Unlike most other disciplines, due to the complexity of their research, fields of science and engineering continue to value the raw computing power accessible through FORTRAN more than the convenience (and built-in inefficiencies) found in other high-level programming languages.
Thank you, John Backus, on behalf of researchers everywhere.