I have said this before, but it is worth repeating. In the era before the 80's, the engine for innovation was academia and research. In the late 80's and 90's, a second leg to this engine came to its own -- industry led innovation. And today, we have a third leg which is equally strong to the other two -- open innovation. Not just "open source" innovation, but "open" -- perhaps the best way to describe it is "community-based."
Take for example IBM's OmniFind Yahoo Edition. It is fantastic, with great reviews. It is built on Lucene, the best open source search engine there is. And IBM has committers contributing back to Lucene. But OYE is closed source. Yet, free to deploy (for a document limit of 500K). And build upon. Or take DB2 Express-C. Free to build, develop, distribute. These are true "open" innovations, because they unleash the power of the community to further innovate. This seamless flow between closed source and open source is what I would call "open" innovation.
I want to emphasize that "closed" source will not improve by simply "absorbing" the best of open source and making it closed. Our clients will see through that. And the communities will close around such closed source predatory practices. That is why, for me, and for IBM, it is a two way street. We need to take advantage of open innovation, but we need to contribute back into the community -- both in closed sources, but equally importantly, in the form of improving the open source projects (e.g. our contributions back to Lucene, and of course the tremedous investment of IBM on Linux).
And I see this model happening in spades around Web 2.0. As I said in my earlier post, it is not about "open" vs. "closed", "new" vs. "old". It is all about mix and match so that the enterprise customers (my focus) benefit from all innovation. It is a false dichotomy to create these polar extremes.