aspirin in Science magazine next month, without patenting, no company will spend $150 million going through Food and Drug Administration approval to commercialize aspirin knowing that the second company will not have to amortize the large investment. In the case of a novel pharmaceutical, one should generally patent inventions to commercialize them. However, if I develop a computer program such as X-Windows, where there is a relatively low cost to take that program from the point of development and get it embedded into products and out into industry, or when there is a desire to make it a standard, maybe the best thing to do is to license it for free. In fact, our license agreement for X-Windows has only one constraint. It costs nothing, by the way, but the one constraint is that you preserve MIT's name on the copyright notice. You can do anything else you want with it.

So I am coming from a slightly different point of view than most of the authors here, and I want to start by saying that I agree with about 90 percent of what John Armstrong says in Chapter 8, but I am going to discuss the 10 percent that I question, or disagree with.

Before I get into that, however, it is to look at some figures—important because they explain the economic incentives for commercializing technology. In other words, if we do not understand the driving motives for commercializing technology, we cannot create policy for intellectual property.

Figure 9-1 gives three scenarios for developing technology. In terms of the investment of money, the first scenario is curve A, which shows a very small investment of money over a long period of time to develop the technology. I call this the minimalist curve, and it is typical, quite frankly, of large U.S. companies. I am going to return to that point, but it is also typical of behavior when intellectual property is not a valuable asset to developing the technology.

The problem with curve A is (1) it gives a long window of opportunity for competition, and (2) the people running these businesses are spending too much time looking for money.

Curve B is the optimum curve in which significant resources are infused early on. If you have heavy protection of intellectual property, it drives your behavior more toward curve B. By the way, one thing this graph does not show is that if the total amount of investment to commercialize a technology is very small, it drives curves A and B closer together. This graph assumes that the total investment to get into the market is very large, which drives curves A and B apart. You can see that over a long period of time, curve B will make a lot more money than curve A. Japanese businesses are curve B companies because they sacrifice short-term profit for long-term gains and market share. Curve C is what happens when you throw too much money at a technology. You can actually corrupt the