Michael Fischer lifts his arms, extends his fingertips, and sweeps across the end of Ingalls Rink in a graceful arc. The 64-year-old computer science professor then returns to his teacher, Olympic ice dancing competitor Melissa Gregory, with a timid smile that betrays his love of the ice. The same smile breaks through his quiet demeanor again over coffee as his fingers trace networks connecting computers in a distributed system.
The Michigan native, who as a child built answering machines out of tape and wire recorders, will mark his 26th year as a Yale professor this semester. He is currently teaching two undergraduate computer science courses while mentoring doctoral candidates in the Computer Science Department.
[ydn-legacy-photo-inline id=”14213″ ]
Fischer talks as though his twin passions for computing and skating were pre-ordained. Just as cold Ann Arbor winters naturally led him to the ice, his father’s position on the mathematics faculty at the University of Michigan inspired Fischer’s fascination with computing.
Fischer notes that the computer science of today looks vastly different than the field he encountered in high school.
“The first PC I saw was this 4-bit processor with a few LEDs,” he said. “My friend had one that made the lights flash in nice ways and he would say, ‘Isn’t this great?’”
Now replaced by sleek, gleaming laptops, the old relics of the 1970s and ’80s gather dust in Fischer’s garage. While Fischer says he “would be happy to part with them” if anyone is interested, one cannot help but detect a trace of nostalgia as he thinks about his collection.
“There’s a lot of history in that,” he said.
It is a history that Fischer has helped shape. His 1985 paper “The Impossibility of Distributed Consensus with One Faulty Process” is considered by many, including Journal of Distributed Computing editor-in-chief and University of Toronto professor Vassos Hadzilacos, to be one of distributed computing’s seminal works.
“I consider this to be the most fundamental result in fault-tolerant distributed computing,” Hadzilacos said in an email.
Hadzilacos said the paper elegantly proves “that there is no algorithm that allows a collection of possibly-faulty computers to reach agreement in a setting that essentially models the Internet.”
According to MIT professor Nancy Lynch, a co-author of the paper and the head of MIT’s Theory of Distributed Systems Research Group, the 1985 research laid the groundwork for systems engineers who want networked computers to continue functioning in the event of one terminal’s corruption. Hadzilacos added that the negative result of the paper led to positive research in the field by scientists trying to develop new ways of network communication in the wake of failure.
The paper earned Fischer, Lynch and University of Warwick professor Michael Paterson the 2001 Edsger W. Dijsktra Prize in distributed computing.
Fischer’s students attest to his capabilities in the classroom. Hong Jiang, who expects to complete his doctorate with Fischer by the end of the year, said Fischer’s kindness and knowledge attracted him to Yale when choosing graduate schools. Jiang, who first met his adviser at a computer science conference in China, said that the breadth of Fischer’s knowledge was compelling.
“No matter what I am interested in at any particular time,” Jiang said, “I can always learn something from him.”
Lynch also cited Fischer as a mentor, saying that, at a time when computer science was largely a masculine field, Fischer was always encouraging of women in the discipline.
“At conferences, it wasn’t unusual to see Mike surrounded by five or six women at the lunch table,” she said. “A lot of the women in the field gravitated toward him.”
Fischer’s current research aims to streamline communication between multiple computers working within a network without discrete addresses. Fischer hopes the research will enhance the functionality of wireless networks.
“It’s like having one computer saying ‘Hey is anybody out there?’ and having a bunch of computers respond, rather than looking for a specific address,” he said.
At the end of almost every day, Fischer dons his skates and hits the rink to practice.
“I’ve skated ever since I was a little kid,” he said.
Still, Fischer has much to learn. Until a few years ago, skating had taken a backseat to research demands. But when Gregory and her partner, Dennis Petukhov, arrived in Connecticut, Fischer “started learning skating all over again.”
“Technique is everything,” Fischer said. “I had progressed about as far as I was going to progress using the wrong techniques.”
Asked about his future, Fischer smiles and shrugs.
“I have no idea.” He says, “Life is full of possibilities.”