Reinventing Discovery - Michael Nielsen [102]
To be successful, grant agencies can’t merely compel openness, they must also forge consent and agreement within the scientific community. If they don’t do this, it’s too easy for scientists to respond by following the letter of grant agency requirements, but not the spirit. Imagine future scientists releasing “open” data sets that are so poorly documented that they’re useless to anyone else. It’s one thing for a scientist to dump raw data online in some obscure location. It’s quite another to carefully document and calibrate that data, to integrate it with other scientists’ data, and to actively encourage other scientists to find new uses for it. That’s what it will take for the scientific data web to succeed. More generally, for networked science to reach its full potential, scientists must make an enthusiastic, wholehearted commitment to new ways of sharing knowledge. For that to happen, grant agencies must work individually with scientific communities, talking at length with scientists in each community about ways that community could become more open. Are there data that could be systematically shared? What about computer code? What about people’s questions and ideas and folk wisdom? What else could be shared? How quickly could it be shared? What new tools need to be developed to make this effective? If the grant agencies do this, they can act as catalysts for Bermuda-style agreements to share scientific knowledge. And, having forged such agreements, they can then express them in policy. This will be long, slow work, but the payoff will be a tremendous cultural shift toward more openness.
Incentivizing Open Science
The prospect of the grant agencies saying “Thou shalt work more openly” leaves me, as a scientist, with mixed feelings. While it will promote the use of new tools, it won’t cause truly enthusiastic adoption of those tools by scientists, unless we also create new incentives to use those tools. Today’s scientists show a relentless drive to write papers because that’s what’s valued by the scientific community. We need new incentives that create a similar drive to share data, code, and other knowledge. How can we make sharing knowledge in new ways just as imperative for scientists as publishing papers is today?
It helps to look at this question in economic terms. In a conventional economy, if I trade you a sofa in return for some cash, you gain a sofa, and I lose a sofa. But scientific discoveries are different. If I share news of a discovery with you, I don’t lose my knowledge of that discovery. This kind of sharing is great for society as a whole, but it has a problem from the point of view of the original discoverer: if they are not recompensed, they have much less reason to invest time and effort to come up with the discovery in the first place.
The solution to this problem adopted by the scientific community in the seventeenth century (and still used today) is brilliant. Instead of giving people exclusive rights to their ideas, as in a conventional economy, we have created an economy based on reputation. Scientists openly share their discoveries by publishing them in scientific papers—essentially, giving them away—but in return they get the right to be credited as the discoverer. By being so credited they can build up a reputation, which can be turned into a paying job. It’s a type of property rights in ideas, leading to an economy based on reputation, and establishing an invisible hand for science that strongly motivates scientists to share their results. The foundation for this reputation economy is a set of very strong social norms: scientists must credit other people’s work; they cannot plagiarize; and scientists judge other scientists’ work by their record of publishpapers. But these norms focus on just one way of sharing scientific knowledge: the scientific paper.