With his usual rigour, Stephen O'Grady considers whether open source is innovative over on his blog. As ever, his view - that "innovation is a function of incentive, not the software development model" - is worth understanding and accepting, but I think there's more to consider here. While it provides no guarantees, I believe an open source environment potentially makes software innovation cheaper and easier.
As a proprietary developer, you are responsible for the eternal care of every line of code you add to your software. In the early days, you can be very productive, creating clean, fresh software that is compelling and doing so fast because you're in complete control of the process. But the code you create is your sole responsibility, and as it gets more and more substantial - and as you have more and more paying customers depending on it - the burden of sustaining it grows.
You may have been responsible for 100% of the innovation, but with proprietary code you're also responsible for 100% of the care and feeding of what you write. That means as you add your next innovation, you're solely responsible for all the ones that came before. That means you're the one responsible for keeping everything working as the operating environment - OS, app server and so on - changes around you. It's like spinning plates.
But in an open source environment, the burden of sustaining is shared among the community. That means that instead of being solely responsible for the sustaining of every innovation they add, innovators can contribute their work to the shared code commons and have the sustaining shared by everyone. They are then liberated to innovate more. Unlike a proprietary code base where the only option is to handle 100% of the sustaining yourself, in an open source project an innovator can expect to have a gradually reducing sustaining burden from each innovation they contribute.
This is also why open source projects make a great starting point for innovators. Freed from the need to re-invent the core, they are able to devise the new more easily, of course. But just as they are no longer responsible for all 100% of the sustaining, they are also able to work alongside other innovators and to benefit from their differing view of the world, bringing new richness. Even proprietary developers who eschew open source depend on this effect (albeit without embracing the benefit of ongoing maintenance of their work by the communities they parasitise). Apple's empire is built on a base of open source, and historically Microsoft has benefited hugely from it.
Whichever way you look at it, the once-common assertion that open source is not innovative is just another spin on the open-source-just-means-no-license-cost frame that proprietary vendors still want us to use. Whether you invalidate it O'Grady's way or mine, it's still the case that open source is proving a fine foundation for software innovation today.