It’s the €6.7 billion (£5.9 billion) project that no one has heard of. The European Open Science Cloud promises to allow researchers, businesses and the public to share and reuse the terabytes of data generated by modern science and the big data revolution.
Rather than leaving their data stored on a memory stick in a drawer, researchers will open it up to their peers across the continent. This should improve the reliability of findings, allow datasets to cross disciplinary boundaries and free up entrepreneurs to create the data-driven digital products of the future.
Or at least that’s the plan. Conceived by the European Union in 2016, the cloud is set to become a reality by 2020, with pilots already running.
But there are fears that progress is too slow and that, with researchers lacking incentives to sign up, it may not have the transformative effect that Brussels hopes for.
“The question is: are we fast enough? I am afraid that we are not,” warned Georg Schütte, state secretary at the German Federal Ministry of Education and Research at the Open Science Conference 2018 in Berlin.
Germany has been building national data and research infrastructure for more than three years, he said, and is only about halfway through.
At the European level, it will perhaps take even longer, Dr Schütte predicted. If only up and running by 2024-25, “the world will have changed fundamentally” and Europe will have moved too slowly to capture the scientific and economic benefits of open data, he told delegates.
To try to understand the project is to struggle through a thicket of working groups, acronyms and organograms. But one thing to grasp is that “cloud” is more a metaphor for a “seamless” commons of data across the continent, as the European Commission puts it – it does not primarily mean the building of new computer infrastructure.
Instead, at least in the early stages of the project, the challenge is to “federate what exists”, according to Jean-Claude Burgelman, head of the Brussels unit in charge of the cloud. “There are 2,000 repositories in Europe and they all want to be part of it,” he said.
Nor is it going to be exclusively European. The cloud is part of a global push towards making data “FAIR”: findable, accessible, interoperable and reusable. This involves making sure data and metadata have such things as unique identifiers and clear reusage licences.
Karel Luyben, the national coordinator for open science in the Netherlands, and one of the driving forces behind FAIR, said that he was talking with research organisations in the US, Australia and some Asian countries. These principles of open data need to reach beyond Europe, “otherwise it’s not going to work”, he warned delegates.
The cloud is part of a wider push for “open science”, the younger cousin of the open-access movement. If open access is about making finished articles publicly accessible, open science focuses on opening up underlying data and methods.
One hope is that this could halt the so-called “reproducibility crisis” in research: by making data accessible, it should become more obvious when scientists have spun their results to get a big headline finding. But why would researchers open themselves up to more scrutiny for no apparent gain?
Scientists are being asked “to act in the interests of science as a whole and the community interest”, said Sarah Jones, an associate director at the Digital Curation Centre, which helps researchers take care of their data. “A lot of those rewards for individuals aren’t there yet.”
Universities need to reward their academics career-wise for making data open, she argued. Open science requirements are likely to become mandatory in the EU’s next research framework from 2021, she added.
There is also a general lack of awareness. According to a survey of researchers last year, less than 15 per cent had heard of the cloud, her presentation said. This is a “really big concern for us”, she explained.
One final hurdle is money. Although costs will vary from discipline to discipline, Professor Luyben estimated that data openness will consume 5 per cent of total research expenses. The commission has estimated that the cloud will cost €6.7 billion of public and private investment, with €2 billion alone set to come out of the Horizon 2020 research programme budget.
So the barriers are sizeable but the costs of failure are much bigger, speakers warned. If the EU does not to build an all-encompassing data cloud quickly, the fear is that private publishers will get there first. “There are lots of very strong players in the market ready to take up the failure of public policy,” said Dr Burgelman.
后记
Print headline: EU is urged not to let open-data cloud drift
