Catching Up with Chuck Johnson
An early innovator shares his insights and vision for the future of mathematicsScientific Computing recently caught up with Charles "Chuck" W. Johnson, an early mathematics innovator, MIT alumni and Corporation Development Committee member, and founder of Visual Numerics (formerly IMSL). During our exchange, Chuck talked about his role in evolving mathematics and statistics into mainstream business analytics, as well as the profound changes the mathematics industry has seen over the past 50 years.
Q: How did you become interested in mathematics?
When I studied civil engineering at the University of Wisconsin in Madison, mathematics was a hobby. At the time, my brother and I used to argue about math at the family dinner table, which made for great conversation. But I didn't pursue mathematics seriously until I went to MIT to earn a graduate degree in civil engineering. As I started using computers there, I became more interested in math and what kinds of math equations I could calculate with computers.
Q: What type of computers did MIT use?
In the early 1950s, we used a five-million-dollar computer called the Whirlwind. It was the fastest computer in the world, with only 2000 bytes of memory, and it occupied 3,300 square feet of a building on MIT's campus. The university and labs used it for engineering and scientific purposes. It was also the prototype of a computer that would be part of a system to track Russian jet bombers and missiles. I thought that was fascinating — it opened my mind.
Q: How did computational computing shape your early career?
During the Korean War, I was a first lieutenant and manager of the computer center at the United States' Army Engineer Research and Development Center (ERDC) in Fort Belvoir, VA, just outside of Washington D.C. Part of my job was to visit manufacturing companies and computer installations and to make recommendations on what the Army should purchase.
When I got out of the Army, I had the right computer experience, so I began working for IBM. My first job was in their applied science group. After that, I merged into the systems engineering group in Milwaukee, WI. I became a troubleshooter and competitive computer expert for IBM, working mostly with their larger customer accounts. As you can imagine, I had to know a lot about IBM's machines and their competitors' machines.
Q: What made you decide to start your own mathematics and statistics company?
In July of 1970, after 12 years of working for IBM, my wife Jen, my friend Ed Battiste and I decided to start a company focused on mathematics and statistics. At the time, IBM had developed a library of mathematical and statistical routines that worked well, but they weren't the company's primary focus. We were impressed with that idea, but didn't think IBM would carry it through.
At that time in the 1960s and 1970s, the computer industry was chaotic. People shared computer programs all the time that weren't always correct or tested. Competitors marketed their software as compatible with other computer systems when they weren't. It was highly unregulated. We knew we could do better. So, we decided to provide routines that were fully tested and compatible across a range of architectures.
Q: Tell us about IMSL. What was significant about those first years developing product?
When Jen, Ed and I started International Mathematical and Statistical Libraries (IMSL) in October 1970, we heard Monsanto and Upjohn say that they would buy a set of tested mathematical and statistical routines if we developed them. So we wasted no time in developing the first library of subroutine.
In July of 1971, our first IMSL Library came out on the IBM computer. It was comprised of 70 percent mathematics and 30 percent statistics, and we sold it to companies who were developing scientific and engineering applications. At one point, we had a government contract to build a language translator that would translate a library to a target computer — what is now called "mapping" or "porting." We were completely ahead of the curve back then.
Q: How did IMSL change the mathematics industry?
During the 1980s, there were a lot of changes in the computer industry. The industry moved from mainframes to minis, workstations and PCs and prices became cheaper. Even though our computational business was strong, we knew that customers wanted visual capabilities, too. So, we acquired two companies in the graphics/visualization business: 3D Vision in Los Angeles, CA, and Precision Visuals in Boulder, CO.
The trade publications loved the idea of a mature computational company getting acquainted with visualization. In fact, we became the first company to merge graphics with a computational library and to offer the combination to the academic and scientific worlds. This was really the advent of visual data analysis. Shortly thereafter, Microsoft introduced Microsoft Office with visualization capabilities at a lower price point for the business world, and it wiped out a lot of smaller visualization companies.
Q: Given the changes you've seen, what motivates you now?
I'm proud that well over 500,000 people use our IMSL Numerical Libraries and visualization software for important research such as forecasting earthquakes, analyzing cancer treatment data, and for hundreds of other critical applications that require math and statistics. I also look at our "Who's Who" of customers in financial, aerospace, automotive and science, and I marvel at them. It makes me feel good that we produce worthwhile products and services for these companies, so that the rest of us can benefit from what they do.
Q: In your perception, what is today's biggest math trend?
High-performance computing has advanced our world's mathematical capabilities to the nth degree. As computing speed doubles every year, so does our ability to develop and use certain libraries and visualization techniques. Recent advancements such as predictive analytics, which require significant speed and mathematics functionality, would have been impossible to build 30 years ago.
Q: What are your predictions for the future of mathematics and statistics over the next 50 years?
I believe that calculations will become even more sophisticated, and mathematics companies will keep adding advanced functionality from other scientific and technical practices. For example, last year Visual Numerics added neural network capabilities to JMSL, our Java-based IMSL library. Neural networks were developed to mimic human problem-solving processes by applying knowledge gained from historical data to new situations, enabling companies to build predictive models that increase in accuracy as they collect new data.
Even though the underpinnings of future mathematical and statistical applications will be sophisticated, we're already making them easier and more intuitive to use. Like the monks who translated the Latin bible in 1500s, I think it's important to relate technical topics to a largely un-technical world, and to educate the masses.
A Few International Mathematical and Statistical Libraries/Visual Numerics Firsts
• launch a United States mathematical and statistical library company
• bind mathematics together with statistics in one comprehensive library
• merge graphics with a computational library
• provide a commercial numerical library in pure Java for mathematical and statistical routines
• introduce neural networks into a computational Java library
• provide a commercial numerical library in pure C# library for .NET applications