Beginnings and LabVIEW
I started working with computers and instrumentation measurements in the late 60s. My first project measured near-field acoustical transducers with multiple elements and predicted the far-field beam pattern, a very measurement-intensive process. We quickly realized that computers could provide an alternative to building dedicated test-and-measurement systems if we combined the measurements with A/D conversion. In 1976 the three of us really dived in, setting up NI and financing it ourselves.
In 1983, I challenged Jeff to create a software package to do for test and measurement what the spreadsheet did for financial analysis. I didn’t give him any hard starting point. I observed that when we tried the experiment at the university, we always began with a block diagram. We knew that instruments had front panels. Jeff incorporated both concepts into what we call VIs (virtual instruments)—elemental building blocks for a programming language.
Jeff discovered Macintosh computers late in 1984. He told me that we needed Macs to run our software, because PCs didn’t have the necessary graphics to do the job. So I gulped real big and said, “Oh, okay.” It turned into a real education. We worked very closely with Apple and learned a lot about platforms and about marketing. We were all in on the introduction of new Macs. The Mac II gave us a computer with sufficient performance for our purposes. That’s when we developed our I/O plug-in boards, because we didn’t have much competition on the Mac. We introduced data-acquisition products, which really jump-started our data-acquisition business. LabVIEW redefined how to build instrumentation systems. Since then we’ve ridden Moore’s Law to where we are today.
In some ways, we were lucky. The Mac guys were both real enthusiasts and customers. Working with them meant that we could get our rather immature product finished before we introduced it on the PC. Only when Windows 3.0 came out in 1990 did it finally offer adequate graphics technology. So we spent October 1986 to September 1992 on the Mac. I investigated alternatives. I looked at OS/2, Unix and others. Fortunately, I waited for Windows.
I often describe NI’s history in terms of decades. The first decade gave us GPIB and LabVIEW. The second decade combined LabVIEW and data acquisition, extending our reach to the measurement side. The third decade produced systems with PXI and CompactRIO. The fourth decade put it all together, combining both tasks and industrial embedded applications and focusing on the platform view. I often compare our platform built around our software—LabVIEW, TestStand and LabWindows, for example— with what Apple has done with iOS and Apple hardware to run literally a million-and-a-half applications. We don’t have quite that many, but we have the same ecosystem of customers and partners around a platform.
Achieving a good design requires a negotiated process, creating a cost-effective product that serves market needs well. It doesn’t depend on an additional part spinning the platform. Any new bits have to fit in and work well with the rest of the components. That approach enables us to make very high-precision measurements—very exciting for production test. And because of our high-speed synchronization, developing a semiconductor test system didn’t require re-architecting. Our existing architecture already gave us the necessary performance for that application.
I started the company to create a job that I wanted, and it was going to work at every point along the curve. From my perspective, it had to work. Otherwise the electronics industry couldn’t grow the way it needed to. We introduced CompactRio in 1994, which brought automation to measurements and measurements to automation. That release set us up for what we call cyber-physical systems, the enabling technology for the IoT [Internet of Things]. Since we’re in the test business, we need to test these things.
I don’t mean to imply that we never did anything that didn’t work. In the mid-90s, we tried to follow a more traditional approach in industrial software. Unfortunately the effort didn’t align well with our platform. If I could do it again, I’d see that we stayed very platform-aligned. We needed to stick to the plan.
Our 100-year plan established a culture of supporting innovation. But “innovation” doesn’t mean that everything works the first time. The first version of LabVIEW was an interpreter. Because that architecture made software development relatively easy, a lot of customers built very complex applications, but under the interpreter they ran much too slowly. We had to compile the code. We strive for innovation with persistence, so you innovate, but you also continuously improve the product.
If we look forward, we expect a long run using our software-based efforts to build instrumentation. Our software gives us the nimbleness, flexibility, adaptability and time-to-market advantages that we need where everything moves at the speed of the internet.
The “NI way” focuses very much on customers and their needs. We look at strategic approaches, helping our customers be successful, identifying our opportunities not only with “important” customers, but every customer. As the company continues to grow, scaling our ability to respond to customers’ needs requires a major effort. I’ve been defining ways to observe and track the process and progress as we move forward. I’m a member of the National Academy of Engineering. The NAE has issued 14 “grand challenges” on how engineers can create better lives and a better future for society. So I’ve measured NI’s success on how we support all 14 of those challenges.
In preparing for my departure, I’ve spent a lot of time showing Alex [Davern] how all the pieces of the puzzle fit together. He works well with the engineers. He’s learned to respect the engineering role as well as his own role on the financial side. Given where we are and the progress of the company, he has the ability to take National Instruments to the next level.
If someone gave me a time machine and I jumped ahead 10 years, what would I want to see at NI? Have we helped the people developing fusion power? Are we testing all the autonomous cars? Have we gotten our product into the Industrial Internet of Things? Are we working with the key players in alternative energy schemes? How are we doing with the NAE’s grand challenges?
Beyond National Instruments, I plan to give back to society more directly rather than through products. To get started, I’m betting that there will be a drug that cures Alzheimer’s. At a minimum, my goal is to identify the successful one, helping either in basic research or somewhere later in the process. That’s project number one. Then there’s what I call “fly by.” I’m looking for projects where somebody has a problem and I help them out.