Kodosky discusses the technology treadmill

National Instruments cofounder Jeff Kodosky chose the NIWeek keynote session today to describe the technology treadmill we are all on and how we can deal with it. Not surprisingly, graphical programming has a role to play, according to Kodosky.

We are massive global experiment involving pumping carbon dioxide into the air, overusing antibiotics, and taking other dubious actions that are making the technology treadmill run faster and faster, he said, adding that if technology got us onto the treadmill, the problems can be addressed though new technologies: fast breeder reactors, for example, that can make use of already produced radioactive waste, or gene sequencing to help develop medications and vaccines.

A key, he said, is to be able to make measurements to develop accurate models of large, expensive structures like wind turbines that can run in real time. Siemens Wind Power Denmark, Kodosky said, implemented such an approach using NI tools

But as we look to technology to solve problems, Kodosky continued, "Another big challenge is facing us. We are not producing enough engineers and scientists. The pipeline of students going into technology is shrinking," despite the fact that, as he put it, the only people who can create the wealth in society and scientists and engineers. We need initiatives like STEM, he said, adding, "We are in a crisis situation; society is slow to recognize it and respond.".

NI, Kodosky said, though enlightened self interest (it wants to educate potential employers and customers) is working to get more engineers in the pipeline. But, he said, engineering education needs to change: "We are still teaching text programming in a multicore world." Learning a text language as a prerequisite for developing multicore programs, he said, is like requiring proficiency at Roman numerals as a prerequisite for algebra.

Kodosky went on to tout the benefits of LabView as the graphical programming environment that can help keep us on the technololgy treadmill as its speed increases. He specifically cited the advantageso f LabView’s parallel data-flow approach vs. text programming’s sequential control-flow approach. He cited as a red herring the view that text programs are more compact and less complex.

"LabView is a real programming language and must be taken seriously," he said, adding that computer science needs to focus on the cyberphysical world, which presents more challenges than programming a desktop computer does. Programming for cyberphysical systems needs to deal with timing connections to real world sensors, for instance. Universities, he added, need to teach embedded RT concepts with multiple targets and physical I/O.

"Design, simulation, and test are all intimately coupled," he said, citing NI’s role in all three areas. "There is no doubt about our succeeding in our endeavor," he said, adding, "We are pulling together at a high level what our customers are already doing."

He described the Web as a key enabling technology, describing an NIWeek wind-turbine example that demonstrated "LabView in a browser" Web application development tools. He described a LabView mashup that combined Google Earth, flikr photos, and US Forest Service data.

He concluded by urging attendees to be role models and inspire the next generation of engineers. (I understand that a video of the address will be posted soon on www.ni.com.)