Paternal pleasure: Interview with Jeff Kodosky, cofounder of National Instruments

Back in 1976, three engineers formed National Instruments, essentially to develop the relationship between scientific instruments and computers. Their first product, a GPIB based interface, was introduced the following year. A decade later, one of those engineers, Jeff Kodosky, made a telling contribution, not just to the company's fortunes, but also to the development of scientific and engineering tools. LabVIEW was a system design tool that simplified test programming by using graphical symbols to represent chunks of code. Kodosky explained the first step was to look beyond automating test. "We were thinking about ways of automating development of measurement programs themselves; that is what led to LabVIEW. A key ingredient was the introduction of the Macintosh computer; that set us on this path of using graphics, not just as a means of displaying data, but also of programming." There were pcs at the time but they didn't have the Mac's graphics capability and they were 16bit machines, while Macs were 32bit. It would be another six years before LabVIEW was deployed on a pc, but this proved to be valuable development time. At the time of the launch of LabVIEW 1, Apple had just sacked Steve Jobs and this didn't help the perception that Macs were little more than toys. Kodosky said: "It was a tough sell, even to our sales people. Most of the market we were focused on – people doing test and measurement – were content to stay with BASIC to write 16bit apps on pcs. But we had people in scientific areas who got the vision of LabVIEW; they saw a tool for them. Of course it was underpowered, but they kept the faith and that encouraged us to work on version 2 that would speed things up." The breakthrough came in 1992, when v2.5 was released for the pc, a platform that was enjoying a meteoric rise in popularity and capability, especially in terms of graphics. "We use the term 'dumb luck' and we had our share of it," admitted Kodosky. "We were fortunate that we were on the Mac, because it was a shielded environment that most were ignoring. We had many years of continuous improvement and innovation in LabVIEW before we finally moved to the pc and it took the world by storm. All of a sudden, everybody noticed it but, by then, it was a mature product." Since those early days of LabVIEW, two major developments have enabled the platform's diversification – real time and fpga. "We heard from customers that LabVIEW was just as applicable to real time control as it was to test and measurement. All along, we had been thinking about instrumentation and waveform acquisition and so on. Control was a different perspective; there was a different concept of what was important – real time response, as opposed to the throughput of the data acquisition. So it took education of our R&D team and sales force, but we built that capability over the years and we are now doing some very sophisticated real time high speed control. Our customers are doing mind boggling things." However, Kodosky believes the fpga will have the biggest impact. "It's almost like fpgas were designed with LabVIEW in mind. Here is a device on which you can construct a circuit and which runs at hardware speed – no drivers, no operating systems. It has reliable, high performance precise timing available and we can map a LabVIEW diagram directly to it. We always had this parallelism in LabVIEW and always felt it was a latent capability that we weren't exploiting. Then multicore processors and fpgas came along and that inherent parallelism in LabVIEW became really powerful – and that is really exciting." Such features have resulted in a host of different ways in which LabVIEW is used. In particular, PXI based systems can be used as platforms for developing design verification tests and for high speed production test. Kodosky describes this as a unification of design and production that works both ways: analysis of production test using LabVIEW can provide useful feedback when it comes to design modifications. "Beyond testing, we are seeing it is being used with our CompactRIO platforms in embedded applications – in intelligent machinery, interconnected and sharing a lot of data. It is exciting to see and we are in the thick of design applications." Asked what will drive the development of LabVIEW, Kodosky answered: "The combination of dual core ARM processors and fpgas, which keep getting bigger and getting more resources, like more multipliers. We have a few guys looking into things like how to accelerate execution on the processors by putting some augmentation of the instruction set into the fpga – we are experimenting with all kinds of things. That combination of multicore processors and fpga is the future and there will be a lot of exciting things coming." When these exciting things do come along, it seems certain that Kodosky will still be involved. While not burdened with operational duties at NI, he remains committed to developing LabVIEW and clearly enjoys it. "It's about the brainstorming and the people I get to work with – lots of really bright folks who are just as motivated on changing the world, making even better products that can make our customers even more productive. It's still a lot of fun and I still get in a lot of trouble for coming home late!" Jeff Kodosky Jeff Kodosky, cofounder of National Instruments, is currently an NI Business and Technology Fellow. Known as 'Father of LabVIEW', his work involves mentoring the company through its development and deployment. Kodosky, who holds 68 patents related to the technology behind LabVIEW, has a passion for education and has been recognised with the Woodrow Wilson Award for Corporate Citizenship for his work with organisations such as The University of Texas at Austin, the Rensselaer Polytechnic Institute and the Austin Lyric Opera.