I don't know much about programming, but I do know a little about competition. I just can't see a company paying big bucks to Big Blue to customize a GPU application and then putting the resulting code in the public domain for its competitors to access for free.
Am I missing something? I see the potentional for a lot of cheating (taking - but not giving back). How will this cheating affect open source development.
The other thing I am having trouble with is the concept of one person or group developing software for free and another company such as Big Blue maintaining and customizing it for profit. It sounds like Big Blue is trying to strip the gravy off of other peoples efforts. This doesn't sound like a concept with a long term future to me.
Am I missing something? I see the potentional for a lot of cheating (taking - but not giving back). How will this cheating affect open source development.
You aren't missing a thing. Open source can work in certain situations, but there are software products that I don't believe will ever be open source. Think of vertical market applications in which there is a limited market and a small number of competitors. There is not enough expertise to achieve a critical mass for an open source solution in the market at large, and there is less than zero incentive for the competitors to collaborate on an open source solution. Plus anyone working for "non-financial" incentives will find those incentives quite limited because of the small size of the community in question.
It's no accident that open source has its biggest success in generic areas like web servers. The need for that functionality is so wide spread that the community is large enough to contain those who will work for free. But even then, open source efforts for the most part simply copy functionality originally available in proprietary systems. They don't really innovate - they just make common stuff available for free. Any attempt at innovation is hampered by the need for consensus for the end result.
As a result, there will also continue to be a need for proprietary systems that do introduce innovations. Things like Web Services and intelligent client systems are being heavily pursued in the proprietary space, and those capabilities will allow entirely new architectures in applications in the next few years. Open source will lag considerably behind in adopting those innovations, and that will allow some companies that use proprietary systems to achieve competitive advantages.
Like others have said, I have absolutely no problem if people want to produce software and give it away for free. (That does not mean it's free in an absolute sense. In some cases, it may take more installation time and configuration time, for example, than proprietary equivalents. Even the head of RedHat said non-technical home users should just use Windows.) But open source advocates should not be religious about it - there will continue to be a need for proprietary systems, and there are spaces in the industry that the open source paradigm simply does not make sense.
If I'm using a product for commercial purposes, my senior management wants the warm feeling of knowing that, if a bug is found which affects production, that some competent organization is going to take responsibility for fixing it NOW, not when somebody happens to feel like it would be a fulfilling experience. Companies WILL pay for the IBM safety net and hand-holding
The neat thing about inventing a new architecture to do something is you can become indespensible for maintaining it, which is one of things the proprietary companies hate (they want to keep their employees in their place). So some of the more successful inventors have grabbed a chunk of power to do the right thing, or cash in, or screw up. But even if they screw up the design or maintenance, other people will step in to take over.