There are many good reasons why we’re told to respect our elders.
With the news that Oracle was cancelling its support for its Fortress project, the issue of old technology clinging to life much longer than expected has once again came to light. Fortress was intended as a replacement of sorts for the legendary Fortran, the first programming language and one that still runs on some of the world’s highest performing computers. While technologically quite different from its predecessor, the name “Fortress” was clearly intended to build on the legacy of one of information technology’s highest achievements.
People pay a lot of attention to the gracefully aging Cobol, programming’s second-oldest profession, but most of us outside of scientific circles rarely hear much about Fortran’s health. In some respects, both are postponing their retirement for the same reason: the “millions of lines of code” that nobody wants to uproot.
But there’s an even more basic reason why computing innovations of the 1950s are still being refined rather than replaced. Developed by IBM, as most new computer innovations were in those days, Fortran was designed as a general-purpose language for scientific calculations. When Cobol came around, a kind of distinction became clear: Fortran would solve big problems (science), Cobol would solve lots of little ones (business).
Today, these problems still exist, and both languages offer viable solutions to them. You can tweak them, you can improve them, but the basic logic they introduced 60 years ago is still valid today.
This kind of evolution is by no means peculiar to information technology. Take the rocket engine as an example. Talk of scramjets and space elevators notwithstanding, there’s still no better technology to get a payload into space than a multiple-stage rocket. While there have been significant advances, there hasn’t been a true revolution in rocketry since the 1950s.
One problem with the IT industry, as I see it, is that the rapid pace of change is seen as an unquestionable good. Since Fortran and Cobol are relics of the past, we should simply forget about them and move on to something better.
Universities that teach courses in Cobol are few and far between, and after speaking to Steve Conway, research vice-president of IDC Inc.’s high-performance computing section, I've heard that Fortran is suffering the same fate.
It’s the newer and flashier kinds of programming that are being taught, he says, most of which cannot perform the scientific computing functions that their ancestor did. A shame, he adds, because high-performance computing was the fastest growing technology market for the 2000s, beating out flat-screen TVs and game consoles. Once again, we’re facing a skills shortage.
New technologies are always exciting. But let’s not have all our brightest minds studying scramjets when we still need plenty of rocket scientists.