Home >  Business

Wednesday, May 15, 2002

By Dan Fitzpatrick, Post-Gazette Staff Writer

Carnegie Mellon University professors Bill Guttman and Bill Scherlis want to make computer software more reliable.

As simple as that sounds, their task is a complicated one. As more people rely on software to bank, shop and do business, the frequency of breakdowns is causing problems that have yet to be fully measured, studied or solved.

Guttman and Scherlis are tackling the dependable computing problem in an unprecedented way, by forming a consortium of university, government and business leaders to make computing systems reliable enough for people to trust with their lives and livelihoods. Not only do they have the nation's No. 1 computer science school behind the effort, dubbed the Sustainable Computing Consortium, but they also have a list of big-name businesses willing to help with the task.

The consortium members, to be announced tomorrow in Washington, D.C., include AIG, Alcoa, Caterpillar, Cisco Systems, CMP Media, Confluence, General Atlantic Partners, Mellon Financial Corp., Microsoft, Pfizer, RedSiren Technologies, Pittsburgh law firm Reed Smith, Tata and UPMC Health System. There are more than 20 founding members, each of which has agreed to pay the consortium $25,000 annually for five years.

The hope is the research among consortium members will help assess the risks and costs of computer failures, which some people suggest may be costing U.S. companies more than $100 billion annually, as well as exploring the possibility of accepted standards for the software industry.

A secondary goal is to turn Pittsburgh into a hub for computer dependability.

"Southwestern Pennsylvania, can become, at minimum, a center of excellence, if not an epicenter, in information network security," said Doug Goodall, president and chief executive officer of computer security firm RedSiren Technologies, one of the founding members of the consortium.

In forming the consortium, CMU is also addressing the vulnerability of this country's computer systems, post Sept. 11. Everything from air traffic control to Pentagon defense systems relies on software to work properly. "By failing to improve the quality of the software we develop," warned a 1999 information technology industry report to President Bill Clinton, "we put the nation at risk."

CMU's decision to create a new consortium to tackle this issue stems from a relationship it has with NASA, which earlier this year tapped CMU to direct a $23.3 million program to improve the space agency's ability to create dependable computer software. In space, the failure of computer software often can be spectacular, such as when NASA lost two Mars probes within months of each other in 1999. The Mars Climate Orbiter burned up when engineers programming its navigation unit confused English measurements for metric ones. Its companion, the Mars Polar Lander, crashed a few months later when its computer shut down its engines before it reached the ground.

Under a five-year agreement with NASA, CMU's School of Computer Science is working with the University of Maryland, the University of Southern California, Massachusetts Institute of Technology, the University of Washington and the University of Wisconsin.

The new program is also raising CMU's profile in Silicon Valley, where the university is renovating several buildings at Moffett Field, Calif., for use as a West Coast branch campus.

For NASA, CMU is studying ways to develop large, complex software systems that are predictably reliable, secure and able to keep running despite component failures. CMU's School of Computer Science Dean Jim Morris explained the basis for the NASA project in an essay published in December 2000.

Morris, in his essay, compared the software industry of the 1990s to the auto industry of the 1950s. "In the 1950s, what was good for the car industry was good for the U.S. . . .As with car quality in the 1950s, it is widely argued that it is a disservice to stockholders to make software more reliable."

"While stories of bad software killing people are still rare, they exist and may portend the future," Morris wrote. He also warned that "we rely upon computers far more than we fully realize until we're forced to step back and take notice of the situation."

The defining quote for the modern computer industry, according to Morris, was a statement made by Silicon Valley pundit Guy Kawasaki in the late 1990s. Kawasaki, referring to the poor quality of some software products, told an audience of entrepreneurs that the ruling philosophy for computer companies is "Don't worry, be crappy."

The line drew a standing ovation.

CMU has been studying computer reliability for decades. Guttman, though, arrived only last year after a career as an expert in the international economic system and as the software entrepreneur who founded Strip District-based printCafe Inc. He took a job as head of CMU's new Software Industry Center and started discussing the problem of computer reliability with CMU professor Ashish Arora and Scherlis, a principal research scientist at CMU who once worked for DARPA, a federal agency. "We felt there was a very significant challenge that wasn't really being addressed, in a direct way, by anybody," Guttman said.

Scherlis said, "What's new here is not the focus on dependability. What is new is the way we are addressing it, which is to bring together the parties who have a stake in dependable systems with technical experts to foster the creation of a new engineering culture."

Several local companies expressed optimism about the effort.

Mark Evans, the president and chief executive officer of Oakland software firm Confluence, said the new consortium presents computer companies with a rare chance to make software more "robust and secure."

Kevin Shearan, senior vice president and head of technology delivery for Mellon Financial Corporation, said CMU's work will help Mellon in the long run. "We cannot deliver financial services without technology," he said. "We must be able to continue assuring our customers that the software engineering practices we have in place provide them with solutions on a consistent, predictable, and high-quality basis."