SchoolTool Calendar 0.10 Released

SchoolTool Calendar 0.10 Released

Let us compare software development to scientific research.

Mathematics or basic science research is very similar to open source software development because information is freely exchanged. In contrast, applied research in private companies where information is only partly shared is similar to proprietary software development. For defense related research, information is not shared at all, except by spies.

Let us estimate with a simplified model the efficiency of different research modes on global scale:

1) Suppose N basic science teams work on related topics in the same field. Each team produces one discovery a year and publishes it with all the details. After one year each team has gained N discoveries, in total N^2 discoveries can be applied in the field.

2) Now suppose N applied science teams do the same, but only publish their result without details. After one year each applied science team can only fully use their own discovery. But they know in part what the other N-1 discoveries are, so may be able to duplicate some of them more easily. So guided they can achieve, say, k times their own results. After one year the field can use kN discoveries. So when k<N, open research is already more efficient globally.

3) Completely secret research is obviously worse. The other teams do not even know which discoveries have been achieved by the others. After one year only N discoveries in total can be used, and perhaps even less if some discoveries have been duplicated by ignorance.

In summary, there is a simple reason why open source software is more efficient when there is a large pool of programmers interested to work on similar projects, it is a multiplicative process: each result is multiplied by the number of participants.
In contrast, proprietary software development is an additive process: the results are proportional to the number of participants. So proprietary software development may make economical sense only when few external experts exist, for niche products. For general purpose software requirring widely spread expertise it is much better for all to adopt an open source development model.