The Geoffrey Boulton's Royal Society report on Science as an Open Enterprise is published today. The headline findings are, unsurprisingly, that science should be more open with its data:
- Scientists need to be more open among themselves and with the public and media
- Greater recognition needs to be given to the value of data gathering, analysis and communication
- Common standards for sharing information are required to make it widely usable
- Publishing data in a reusable form to support findings must be mandatory
- More experts in managing and supporting the use of digital data are required
- New software tools need to be developed to analyse the growing amount of data being gathered.
This all sounds well and good, but of course the devil is in the detail. I wondered about this next bit for example, part of the section on computer simulations:
In principle, simulations whose conclusions have major implications for society should be assessable by citizens. However, simulations such as those that evaluate the operation of cities as a basis for planning or that simulate the operation of climate and environmental systems and forecast their futures, are often highly complex, and only truly assessable by experts. Nonetheless, it is important that the operation and output of simulations where major public interest is at stake should be set out in ways that expose problematic issues and identify levels of uncertainty both in their forecasts and in their implications for policy. Efforts of bodies such as the Climate Code Foundation in improving the transparency and public communication of the software used in climate science are important in this regard.
The words "in principle" seem important. Boulton is suggesting that where a computer model is big and complex enough, the public are just going to have to accept the scientists' word that they are right. Since government-funded academics are always going to have access to more computing power than Joe Public, it will therefore always be impossible to show that published results of such simulations are actually produced by the code.
One can reasonably wonder whether public policy should be informed by any research for which the data and code are not freely available, but it seems to me that it might be possible to sidestep the issue: the public could demand, as an minimum, that such simulations demonstrate predictive power - forecasting, not hindcasting - before they are adopted as tools of public policy. The alternative is to make scientists legally liable for their professional advice, in the same way as engineers or other professionals are.
Unfortunately, Boulton's report does not consider issues such as these, but in any case, there is no reason that the computer code underlying large computer simulations should not be made available, and Boulton's bleating about complexity seems like a smokescreen to me - complexity is usually not a problem for crowdsourcing efforts.
Boulton also raises the spectre of intellectual property rights. This is, frankly, an insulting argument. If global warming is the existential threat that Boulton and others involved in the climate field have said it is, they should have no objections to waiving any expectations of IPR. To suggest otherwise seems extremely revealing to me. As someone said the other day, on the day when alarmist climatologists start to look as if they are taking global warming seriously then maybe the rest of us will.
Digging further, you find that the report recommends placing data in online repositories "where data justify it". Reasonable people would be justified in being worried. However, there is also this recommendation:
As a condition of publication, scientific journals should enforce a requirement that the data on which the argument of the article depends should be accessible, assessable, usable and traceable through information in the article. This should be in line with the practical limits for that field of research. The article should indicate when and under what conditions the data will be available for others to access.
This sounds much more solid, but again I have my doubts. In the wake of Climategate, I contacted the Committee on Publication Ethics, an umbrella body for scientific journals, suggesting that they try to formulate a policy on data availability (as well as one on inappropriate approaches to journal editors) for their members. However, two and a half years later, and despite several reminders, they have yet to get round to considering the issue. I will not therefore hold my breath waiting for journals to deal with the data availability issue.
The report also mentions Climategate at one point, in terms that are largely unobjectionable:
The potential loss of trust in the scientific enterprise through failure to recognise the legitimate public interest in scientific information was painfully exemplified in the furore surrounding the improper release of emails from the University of East Anglia. These emails suggested systematic attempts to prevent access to data about one of the great global issues of the day - climate change. The researchers had failed to respond to repeated requests for sight of the data underpinning their publications, so that those seeking data had no recourse other than to use the Freedom of Information Act (FoIA) to request that the data be released.
The report seems to contain many pious words, but raises as many questions as it answers. I have little sense that it will bring about any great change.