figure 8
So what I think is the most important part, at least for me, aside from the fact that it's a lot more fun to develop this way, is that the open approach is so powerful because we can leverage all the parts of evolution, not just the mutation or incremental change, but also mutations in parallel, thanks to the open exchange of the genes of software, the source code. Also, in the open source community there are often, but not always, requirements on being able to recombine. The General Public License, authored by Richard Stallman, which is what Linux uses as a license, for example, has the explicit requirement that you are allowed to recombine your source code with other people's source code. You cannot take that right away, and that's a very important thing. It doesn't always have to be in the license itself. There are open source products that try to encourage the combination not through licenses but, for example, through social customs. This is the one that science uses. The thing that makes scientists tell other scientists how to reproduce and how to improve on their work is the social customs in the science world. There are no licensees in science but there are very strong social customs. Over hundreds of years we have evolved a system that encourages the recombination you find in science. I already mentioned the parallelism by making source code free. Another thing that a lot of open source advocates, like Richard Stallman, talk about is trying to avoid things like patents, which limit people who can do certain things, and by limiting it to one company or a few companies, it also limits the parallelism of having people everywhere work on a project. An open source encourages competition. It's very strange to me how, especially in the United States, which is supposed to be the land that glorifies competition, at the same time people tend to glorify a lot of anti-competitive measures, such as patents. Intellectual property rights are all inherently anti-competitive, and that has always been their whole point. Open Source tries to make competition a good thing. For example, in my personal project, Linux, we almost always have not just my source tree but there at least three or four other source trees maintained by different Linux distributors who take other changes than just the ones I create, and actually maintain competition even within the one project. This is not even to mention the whole ecosystem where they take different parts of other Open Source programs and try to compete within the whole ecosystem, too. Figure 8 I will talk about some problems, because there is no question that the open source approach also has some downsides to it, and neglecting those downsides would not be entirely fair. One effect of trying to make things more parallel is that it usually is duplicating a lot of effort. You have parallel projects that expend a lot of energy. There is wasted effort from people working on the same thing. Especially if it's something really interesting, you will find that a lot of different groups want to work in the same area. This is required in order to get that effect of competition. But at the same time, it means that this is very hard to do within a company, where trying to maintain multiple tracks and trying to create an Open Source model within a company might not be economically viable. And I think it actually is economically viable only if you allow it to go outside a company. So you can't just cherry-pick the good sides of Open Source and try to apply it within your company, because you will inevitably lose the economic possibility of having multiple tracks. If you have a wide base, you can spread the cost out, which is how, for example, biology does it, where it's very expensive on a global scale to have millions and billions and the individuals reproducing, but on an individual scale it's actually a lot of fun.