The inventor and leader of the first implementation team of the Fortran language has passed away [1].

I will never forget his thrilling account of how he and his team developed the first Fortran compiler for the IBM 701 in 1954, against the objections of assembler programmers that a machine could never generate efficient code.

As it happened, to map variables onto the small number of registers, Backus and his team discovered that a statistical technique called Monte Carlo worked well, and in using this technique, anticipated compiler optimization theory by at least 20 years.

Fortran was my first high-level language, and to use it, I had to debug the IBM 1401 8K runnable compiler on 2000+ cards. This compiler, unlike Backus', used the Purdue University approach which was to generate interpretive code.

We shall all miss John Backus, and if there is a better land up above, we all I am certain hope he gets to GO TO that land FOR I = NOW,ALLTIME.

Gee, I hope the above compiles. I haven't used Fortran since the 1980s. It certainly is lame.

Thanks for all the hard work, John Backus.