"Innovation is not the same as invention--there's distinction," said Paul Horn, IBM senior VP of IBM Research at the Emerging Technologies Conference put on by the university's Technology Review magazine.
"Innovation is not just great ideas but channels into the market," he said. "Success in R&D requires rapid flow of innovation into the marketplace. Time to market is critical."
IBM ensures that its R&D efforts fill a market or customer need by having its developers spend time with customers "to understand their problems, he said. As an example, he cited the BlueGene/L supercomputer, which IBM this week said ran a benchmark that makes it the world's fastest. The computer draws 1/28th the power of the previous fastest supercomputer, NEC's Earth Simulator in Japan. "That's a good example of an exploratory project to fundamentally change how we think about computing," Horn said.
"Exploratory work can survive, thrive in a world where there's a balance between short term and long term," he said. He told attendees to "get your researchers out in the real world and talk to customers [about innovations] that can change the world, not just their own little labs."
Standards are vital to innovation as well. "Open standards, open source are critical for speedy innovation at a company, in an ecosystem, in the country," he said. IBM holds more patents than any other company; however, patents that can potentially set standards should be made available at little cost to others, he said.
Working with universities is becoming increasingly difficult for large labs, Horn said. Delays, mainly because of concerns about ownership of intellectual property, hinder development work. Some professors "see the potential to get rich with IP," and that, too, delays R&D projects. Said Horn, "There's nothing worse than being slow."