Computer scientist Kathleen Fisher, whose latest project seeks to protect motor vehicles from hacking, says too few people understand how fascinating her field is
In the field of computer science, there is a classic quandary known as the “halting problem.” Devised by computing pioneer Alan Turing in 1936, the key to the halting problem is that it is impossible to solve.
It was the halting problem—which, briefly, attempts to figure out whether any particular computer program will finish running, or run forever—that jumpstarted the career of computer science professor Kathleen Fisher.
She had arrived as an undergraduate at Stanford in the late 1980s, intending to major in physics and lugging with her an IBM PS/2 Model 60 with a 100-megabyte hard drive—a pretty fearsome machine for its day.
“Like pretty much almost all freshmen women, I had no knowledge of, or interest in, computer science as a career,” she says. “But I was somewhat embarrassed that I had this powerful computer, and that I didn’t know how to use it.” So she took a couple of comp sci classes. In talking with her instructor one day, he brought up the halting problem.
“It was the coolest thing I’d ever seen,” Fisher says. “I switched my major to computer science on the spot.”
Since then, her career path has run through industry, government research and academia. Fisher spent 15 years at AT&T Labs Research, renowned for its inventions related to networking and communications, but says she long imagined she would end up in academia. When her daughter headed off to college, she decided it was time, and so she joined Tufts in 2011.
But then another opportunity quickly arose—a three-year stint as a program manager for the Defense Advanced Research Projects Agency (DARPA), the federal agency that tackles some of the trickiest research problems for the military. The agency's work has trickled down to some of the most important technology in the civilian sector.
She came back to Tufts in September, beginning a role she refers to as a “dream deferred”—professor of computer science in the School of Engineering.
Her teaching and research specialty is programming language design, with a focus on what’s called domain-specific languages. Basically, that’s adapting general-purpose programming languages—the most well-known ones include C++ and Java—for specialty areas.
She and associate professor Samuel Guyer are collaborating on a project to improve the tools and structure for building these domain-specific languages—essentially software that is embedded in everything from semiconductor fabrication plants to refrigerators and cars.
Warding off Car Hackers
When she was at DARPA, where teams of researchers are brought on board to work on portions of specific projects, she worked on two areas in her specialty: one involved creating programming languages related to machine-learning applications—basically how computers can learn from large amounts of data—while the other focused on ways to make motor vehicles less vulnerable to hacking.
Modern cars, after all, are pretty much computers on wheels, with vital functions that rely on software: antilock brakes, for example, or cruise control. The typical late-model car has anywhere from 30 to 100 embedded control units, and all cars sold in the U.S. since 1996 are required to have diagnostic ports under the steering wheel through which the vehicle’s “brain” can be accessed.
Having such heavily computerized vehicles makes life easier for drivers. It also means that thieves can now steal cars by hacking into them—no more smashing a window and hot-wiring the ignition. (The London Metropolitan Police, for instance, reported that half of all autos stolen in London in 2013 were entered wirelessly.) And it means that someone with malicious intent could gain control of a car without ever touching it—by infecting it at a repair shop or hacking the Bluetooth or roadside assistance systems, or even by using malware hidden on a CD played on the stereo system.
Fisher’s DARPA project—known by the acronym HACMS, for High Assurance Cyber Military Systems—was devoted to finding ways to build vehicular software that would be invulnerable to outside attacks. Of course, the Department of Defense’s first priority wasn’t safeguarding our Subarus and Buicks, but rather the thousands of military vehicles—on land, sea and air—and drones that rely on similar technologies.
It’s easy to see how the fruits of this work will affect more than military endeavors, in the same way, for example, that the development of voice recognition software by DARPA paved the way for the iPhone’s Siri and other popular consumer products.
A Supportive Field
Being a woman in computer science, Fisher acknowledges, puts her in the minority. “I don’t mind being an outlier,” she says. “I played Little League baseball for six years, and was captain of the men’s golf team in high school.”
Being a woman in a field dominated by men has not hampered her career. “My peers and my research community are very supportive,” she says. “One of the challenges is making a name for yourself. Being one of the few women in the room, you stand out the minute you walk in the door. You have to have good ideas, of course, but people are already paying attention.”
Fisher points out that many of the early leaders in computing were women. (Arguably the most famous was Grace Hopper, a rear admiral in the Navy whose work led to the development of the COBOL programming language and who coined the term “debugging.”)
Nowadays, however, there is a clear gender imbalance in the field. According to the federal Department of Labor, women made up 26 percent of the computing workforce in 2013. In 2012, 18 percent of those who received bachelor’s degrees in computer science were female—a significant drop from 1985, when women made up 37 percent of computer science degree recipients.
One key to attracting more women into computer science, Fisher says, is exposure. “Most women don’t get any chance to learn what computer science is. Why would you choose to go into a field when you have no idea what it is? Most people think computer science means knowing how to use PowerPoint and Excel, or they think it’s about IT,” says Fisher, who has been a co-chair of the Computing Research Association’s Committee on the Status of Women. “So exposure is the key issue.”
In high schools, she says, the subject is too often dominated by a subculture of computer geeks—“guys who have taught themselves to program and who derive status from being the ones in the know and actively push others away. It has a chilling effect on everyone else—women, minorities, other guys who are not in the club,” she says.
To help solve the gender imbalance, Fisher says, a basic understanding of computer science should be required of all students, either at the high school or college level. “It’s a good idea for society to have a good appreciation of the fundamentals of computing,” she says. “If everybody had that, some reasonable fraction of women would discover they liked it and are good at it and would go into the field.”
Helene Ragovin can be reached at helene.ragovin@tufts.edu.