Fortran developer’s obit: at last, some mainstream recognition

John W. Backus, who conceived of and led the IBM team that developed Fortran and released it in 1957, died recently. What struck me about his passing was not so much his (and his team's) accomplishments, but the simple fact that his passing, at age 82, merited some mention in the mainstream press such as the New York Times. (Of course, EE Times covered it, but you would expect that.)

In addition to Fortran, he was heavily involved in development of Algol (for publishing algorithms) and Backus-Naur notation, (for defining a formal language syntax). While much of his work and concepts seem like old news to us, at the time they were radical ventures into unknown, uncharted territory.

Scientists and engineers simply don't get much recognition or respect in the media, except when something goes wrong and some quick “expert” analysis is needed. The real work of the creative process, the struggle against both understood and mystifying problems, challenging accepted approaches, and pursuing dead ends until success is (hopefully) achieved) is ignored. If an engineer or scientist is recognized by the non-technical press upon death, that's a small victory; rarely is the individual recognized outside the technical community while alive.

On the other hand, if you are an artist or film maker or writer, especially one who “pushes the envelope” and offends sensibilities, you get lots of coverage. After all, you're doing something which is smoothly referred to as “exciting” or “dramatic.”

To me, this situation reflects, and contributes to, the view that science and especially engineering is both dull and predictable. As I have written on previous occasions (click here and here), the Intel-inspired public promotion of the semiconductor-process road map was, in my view, one of the worst things to ever happen to our profession. It made process and product development seems routine and predictable, when we all know they are not. And the irony is that it is a self-inflicted wound.

John Backus was, according to the obituaries, a maverick, a “lost soul” as a student (he dropped out of several schools), but he found his niche as a mathematics graduate student. He was hired by IBM while he was on a open-to-the-public tour of IBM headquarters in New York, when he mentioned he was a math student. Nothing predictable or conventional there! (Try to get hired today on the spot, just on the say-so of a project manager, without a formal resume, or involvement of the HR department.)

As for Fortran, it's hard for us today, now 50+ years into the computer world, to understand what a conceptual breakthrough it was, and how difficult it was to actually implement. I have a personal soft spot for the language. Like many students, I was did my early programming in machine language and assembly language, as part of the learning process (some called it trial by fire). But when I finally got to the high-level language (HLL) world of Fortran, I felt like I was liberated. And when I found that there was an even easier language (BASIC), I really started to fly. {Purists will remind me that Fortran and BASIC were very different in concept, with one compiled, the other interpreted, but that's not relevant to my point here.]

Even today, I keep a copy of TrueBASIC on my PC ($40)). It's very handy for quick programs when I need to work out a personal problem such how many permutations there are of a certain set, or where a maximum is in an equation.

Perhaps John Backus got the last laugh. When we faced the Y2K challenge (and yes, it was a large potential problem, despite the fact that the eventual impact was relatively slight), one problem in our industry was getting enough of programmers who could handle the dinosaur of languages. That Fortran lasted as long as it has, and that the HLL concepts it embodies are still viable, is a testament to its brilliance.

0 comments on “Fortran developer’s obit: at last, some mainstream recognition

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.