Rebecca Parsons, CTO of ThoughtWorks; photo by Martin Fowler.
“There is a common perception that once you go to work, you never go back to school full time. But that is exactly what I did,” said Rebecca Parsons, ThoughtWorks Chief Technology Officer and Rice University Computer Science alumna (Ph.D. ’92).
Parsons finished her undergraduate degree and worked in the technology industry for seven years, pursuing a master’s degree around her full-time work schedule. In 1987, she traded full-time employment for full-time graduate school.
“You can, in fact, go back to the classroom. It was quite a transition, less so because I’d taken those master’s classes. But I still felt a twinge of panic, wondering if I still remembered how to take notes or pass an exam.
“Still, the Ph.D. program provided the headspace I needed to go very deep in a particular topic that was also very broad, and I wanted to be able to study and understand it at both levels. I’d already realized I couldn’t accomplish that type of focused research while working full-time.”
Parsons was originally drawn to the architecture of computers when she was first exposed to mainframes and data punch cards as a teenager. The physical connection between the cards, computer, and results intrigued her.
She said, “There was a deeper connection to the actual hardware at that time and you had to think a lot more about what made it tick. That got me interested. I knew what assembler language looked like and I started wondering what it would take to translate from BASIC into something that worked on the machine.”
The co-op program between Caterpillar Tractor Company and Bradley University led to part-time and full-time employment for Parsons. She worked in the manufacturer’s warehouse, helping retrofit one of the first warehouses built in the United States in preparation for automation.
“Working at Caterpillar during my breaks and summers helped me see the application of what I was learning in computer science classes. I was writing in assembly language and communication drivers to transfer data between an IBM mainframe and pdp11 minicomputer.”
“I was also working on control software. When a loaded pallet moving through the warehouse got to a certain intersection, did it need to turn left or right? I was writing those kinds of communications for machine decision-making.”
Her second job took her into a modern warehouse, but she retained the earlier architecture. “It was for low-level tasks, but it was beautiful architecture and could be easily transferred to the modern warehouse equipment. I appreciated the architectural characteristics and how getting something right makes it more portable.”
Parsons said she’s always been a language person, perhaps because she sits on the fence between theory and practice.
“I like to see things through, experience the application of the science. I also loved the theory courses and building expertise in different kinds of languages. But the practical side of me was captivated by Ken Kennedy’s work on compilers and distributed computers. How could we make it easier to take advantage of the existing hardware?”
Her unique combination of theoretical concepts and the practical applications to make computers easier to use efficiently found a home in the Los Alamos National Laboratory, where she worked as a postdoc for two years.
“At the time, there were applications and simulations that took 27 hours to run. People would say, ‘You only pay a 10% performance penalty for going to C++’ but tagging on another two or three hours didn’t convince many physicists to go from FORTRAN to C++. That’s where the value of compilers became evident.”
These days, Parsons is exploring two different trends in technology: evolutionary architecture and systems that mix hardware and software solutions.
“I’ve co-authored a book about building evolutionary architectures where we extend agile software development principles to the way we think about architecture in the technology stack and the overall integration scheme between various parts of the system.
“Architecture used to mean the hardware chip when I was in school, but I’ve always approached systems from the software perspective. Now, I’m looking through that lens at hardware. Roadmaps are changing so quickly now, how can organizations move away from their five- or six-year-old architectures to stay current?”
Parsons used the iPhone as an example, noting dramatic shifts in the way consumers experience and utilize technology in the eleven years since its launch. In turn, other technology providers have been impacted by the demand to keep up with the changing landscape and customer expectations regarding the software they are producing.
She said, “More and more millennials and Gen Z employees are entering the workforce. Their expectations are linked to consumer products, and they have no patience for clunky, internal systems. Our corporate, internal systems and the way they are governed– that too all has to change.”
Parsons also joked about getting back to her warehouse roots with systems that merge hardware and software solutions. “These kinds of systems are gaining traction in the manufacturing space. Evolving machine learning/AI sensors for use in factories is really big. How can we make better use of sensors, intelligent forklifts, or cars to make a physical thing adapt better to its environment?”
Although she’s fascinated by the technology industry’s changing landscape, what Parsons values most is time. Communicating with her many different stakeholders — from the press and analysts to clients and colleagues in different parts of the ThoughtWorks organization — requires her focused attention.
“I crave having enough time to just sit down and write as I think through a problem. I really enjoyed my first sabbatical, when I could explore the first, second, and third order of consequences to proposed solutions. I like having time to let things develop. It’s great to relax and think about what I want to say and how to say it.”