SHARING AMERICA'S TECH NEWS FROM THE VALLEY TO THE ALLEY
by Jeremy Jacquot, courtesy ArsTechnica
Private sector involvement could help make research data publicly accessible. –
While some in science continue to grumble in private about the need to share their findings with the wider public, there is now widespread—if not overly enthusiastic—acceptance among the scientific community that their work should be more transparent. Not that they may have much of a choice in the matter soon if the US Office of Science and Technology Policy (OSTP) gets its way. Earlier this year, the OSTP issued a widely circulated memo in which it called on all federal agencies with R&D budgets of over $100 million to come up with strategies to make their research data more publicly accessible by this September.
The catch: they need to do so without the benefit of additional money, setting up a chain of potentially untenable situations in which research funds would have to be sacrificed at the expense of data management funds, or worse.
In a Policy Forum published in today’s issue of Science, Francine Berman, a professor of computer science at Rensselaer Polytechnic Institute, and Google Chief Internet Evangelist Vint Cerf argue that only by fostering greater private-public sector partnerships can federal agencies and researchers ensure that their data is more widely accessible and preserved for posterity. They start off by citing a few examples of successful public and private databases but go on to caution that most research data is at risk of becoming lost or “homeless” without better management. And they soberly conclude that there is no “magic bullet” that “does not require someone, somewhere, to pay.”
To entice private sector actors like, say, Google to assume digital stewardship of massive research databases, Berman and Cerf stress the need for federal and state governments to offer direct financial incentives such as tax breaks. But governments also need to establish a viable succession plan should the companies decide to move on. Governments should also be willing to provide the initial seed capital for promising new private-public partnerships with potentially viable revenue models and should be clear about which databases they are committed to supporting (and for how long) to ensure that those that need funding receive it and those that don’t remain useful repositories of knowledge.
For their part, scientists should either be ready to open their wallets to access journal articles and databases on a per-item basis à la iTunes or Amazon or be prepared to sit through a series of online ads to access them freely. “Imagine supporting the National Virtual Observatory (astronomy data) from telescope advertisements or paying a small download fee for data from digital marine collections in the same way we download music from the Internet,” they suggest.
For better or worse, the latter approach seems unlikely to fly among beleaguered scientists who have already seen their grant funding whittled away in recent years. But the others provide a plausible roadmap for how databases could be sustainably managed in the near future.
Expecting researchers or funding agencies to willingly make further cuts in their budgets to develop and continually update ever-growing data repositories is at best naïve and at worst disingenuous. Although some will blanch at the notion of adopting a more entrepreneurial-minded approach to storing and distributing data, the sad reality is that until Congress decides to stanch the bleeding (unlikely) or increase the amount of funding (even less likely), there are few other viable options.
Ultimately, these approaches will succeed or fail on the robustness and community-wide appeal of the databases they help create and maintain. The potential downside to getting many participants—public or private—involved is that you risk ending up with a multitude of incomplete databases that impede the progress of research by balkanizing, rather than unifying, disparate datasets. Then again, that still seems like a preferable scenario to having no publicly available datasets at all.
Thank you, TiA