The openness revolution will not be trivial - opinion - 21 June 2012 - New Scientist

“On Tuesday the UK government's Finch Committee came out strongly in favour of making research results freely available to all. Today the Royal Society releases its long-awaited report on all aspects of openness, Science as an open enterprise... As expected, the broad thrust of the Royal Society's report is that openness is a good thing and that there ought to be more of it. Scientists should routinely release their data and methods; universities and funders should support and reward those who do; even industry should be more open with its data... One of the motivations for the Royal Society's report was the so-called "Climategate" episode of November 2009, where hundreds of climate science emails were hacked and leaked. Subsequent inquiries cleared the scientists involved of any wrongdoing, but encouraged them to consider how they might communicate their results more openly and transparently. So as a climate scientist I can offer a personal perspective on the Royal Society's recommendations. Broadly speaking, they are to be welcomed... Meteorology and climate science generate vast amounts of data from weather stations, balloons, aeroplanes and satellites across the globe, and computer models. These are generally very well curated by the meteorological agencies, and fairly easily accessible. However, particular problems can arise with data from individuals or small groups of researchers. The handling of these data has, until recently, been somewhat ad hoc, with only colleagues of the data ‘owner’ likely to have access to them. This is not a desirable state of affairs. But maintaining an archive is not trivial. The data have to be searchable and easily retrievable... Another problem is that academics are under increasing pressure from funders to produce trailblazing research or results of high societal impact. This militates against the more mundane demands of data curation. An issue not addressed by the report is whether the computer programmes used to turn data into results and conclusions should also be made open. Such programmes range from simple number-crunching packages to hugely complex models that have taken millions of hours of labour to produce... Finally, a problem with all data analysis is how to present statistical uncertainty. This is particularly important in climate science... The desire to communicate the findings of climate science more effectively in open and transparent ways has led the UK's Royal Meteorological Society to set up a Climate Science Communications project. It has also launched an open access journal, Geoscience Data Journal, which publishes peer-reviewed scientific data in a way that allows the reader to fully understand when, how and why the data were collected. It is my sincere hope that these sorts of activities will contribute to addressing the problem.”