The only caveat I have to put on it is that you can't test the counter-factual way—that is, if you didn't do this, what would have happened otherwise? I have no way to test it. You can certainly look to the companies and the stream of technology to see where it was inserted into a product and to see how those products have turned out in terms of whether they're a commercial success, but you can't test what would have happened otherwise.
I just offer that as a challenge. It's not an obfuscation; it's more or less an issue in this area of programming generally. The overall theory here is that if you don't provide some incentive to the highly risky speculative activity that is the early stage of R and D effort, you would get less of it than is socially desirable; therefore, there is support and incentive in different forms on the part of the public.