The Programmer's Future

Offshore outsourcing emphasizes how risky it is to be purely a computer specialist, Morris says. Carnegie Mellon teaches the foundation of programming and how to learn new languages. Yet Morris compares computer science for science-focused kids to an English major for liberal-arts minded students; it's a foundation to pursue any career, from science or medicine to business or public policy. "We'll teach you enough to get a summer job in programming, but our deeper purpose is to teach you everything you need to know about digital technology," he says.

So how important will programming skills be for business-technology workers and executives of the future? Clearly, business knowledge--and in many cases industry-specific knowledge--has never been more vital.

Years before Y2K and the popularity of enterprise apps, management-consulting firm Booz Allen Hamilton's IT group shifted its primary focus from coding to business analysis. All of Booz Allen's business analysts were once programmers, but what separates them from by-the-hour labor is their company knowledge. "If we need coders, we can use contractors," says George Tillman, Booz Allen's CIO. "But what we cannot buy is the company knowledge within the business analyst."

Al Biland, CIO at Snap-on Inc., says programming remains a viable route to an executive IT role, but it will compete with other job experience, such as business-unit manager. Tillman predicts it will be more likely for a CIO to come from the business-unit ranks. "As technology changes--the hardware as well as how it's delivered to the user--the need for a technical background diminishes," says Tillman, who started his career as a programmer. PolyOne's Smith, too, sees business analysts as potential CIOs because of their deep understanding of business, strong appreciation for technology, and ability to lead projects and teams. But Archipelago's Mueller says programmers are so often tied to their keyboards, few have opportunities to make the customer or even internal contacts needed to have a far-reaching impact on their companies or move into the executive ranks.

It's tough to imagine a programmer having a greater impact--or a more successful career--than Charles Simonyi. He joined Microsoft in 1981 when it had 40 employees and led the programmers who created Word and Excel, earning him the fortune that lets him fund a software startup he now runs. Simonyi doesn't offer much hope for maintaining today's U.S. programmer numbers. Where he sees hope is beyond outsourcing, in what could truly reshape programming: mechanization.

"To me, the outsourcing trend indicates that an ever-larger part of IT work has become routine, repetitive, and low-bandwidth; one might even say unexciting or boring," Simonyi says in an E-mail. Ultra-educated Indian programmers are underutilized, just as programmers have been in this country, he says. He has a stake in this vision, since his company, Intentional Software Corp., develops tools to make it easier to capture the design of software in the actual code. Yet he predicts a long and painful journey. In the near term, the 20% or more least-productive U.S. programmers could lose their jobs to overseas employees. But those jobs will eventually be mechanized, by utilizing the programming skills of senior U.S. people, Simonyi predicts.

If routine IT programming can be mechanized, Simonyi says, it will bring Moore's Law to software--a steady increase in capacity along with a steady decrease in cost. The basics of software engineering will remain important to creating software, but the focus will be automating functions. Languages and compilers will have the limited, narrow role that order codes and assemblers have today, he says. And obsolescence will be a constant threat. "A programmer's 3-year-old experience will be like a 3-year-old laptop is today: a quarter of capacity, a quarter of speed, and ready for replacement," he says.

It's hardly a comforting vision. But high-tech has never been about comfort, especially when it comes to change. "Moore's Law isn't predicated on keeping the key parameters and key technologies constant," Simonyi says. "We on the software side should take a hint from this."