The US Department of Energy (DOE) has revealed today how papers from research it funds will become free to read, making it the first federal agency to respond to new standards for open access and data-sharing ordered by the White House 18 months ago.
The plans mean the DOE will be releasing up to 30,000 papers annually from behind paywalls, although the directive from the White House’s Office of Science and Technology Policy (OSTP) says that papers need not be made free until a year after publication. The plans come a week after the DOE’s announcement that its researchers should openly share the data from papers they publish.
Open-access advocates have welcomed the plan but say that it is vague and disappointing on some key points. For example, it seems that ‘free’ manuscripts may not be legally open to bulk downloading, re-distribution and re-use for creative purposes such as text mining, even though the OSTP directive had hinted otherwise.
“The DOE’s plan takes steps towards achieving the goals of the directive, but falls short in some key areas,” says Heather Joseph, executive director of the Scholarly Publishing and Academic Resources Coalition (SPARC) in Washington DC. “We don’t want to end up in a ‘read-only’ world of US science articles,” she adds.
Much of the debate following the White House order had centred on whether agencies would create online stores of their free papers, or would link out to free papers elsewhere. The US National Institutes of Health (NIH), the only other federal agency with a public-access policy, had created the popular PubMed Central (PMC), a store of free-to-read manuscripts. But publishers argued this repository was taking away their online hits, and instead launched their own effort, called CHORUS, by which they proposed to make the final published version (or ‘version of record’) of US government-funded papers free at their own websites.
The DOE has decided not to copy the NIH, but is instead linking out to free papers from a centralized web-portal called PAGES, which goes live today. The agency wants two versions of a paper to be freely accessible; it is working with CHORUS so that the version of record will be free on publishers’ websites, but it also wants its researchers to archive copies of accepted manuscripts in online repositories. Exactly when the DOE’s researchers will have to upload their manuscripts online isn’t immediately clear, however.
The DOE will link out to these papers from its PAGES portal 12 months after publication, says Brian Hitson, from the agency’s Office of Scientific and Technical Information, with the aim of linking to the publisher’s version of record when it is available. The agency will also gather the accepted manuscripts together to create a ‘dark archive’ that it can release should publishers’ websites or repositories go down. Hitson said that the agency didn’t yet know whether it might place sanctions on researchers who didn’t comply with its manuscript archiving policy, as the NIH is beginning to do.
Roilings over ‘re-use’
Open-access advocates said the agency is ducking one of the most contentious and poorly-defined issues in the OSTP directive: while the papers will be free to read, the DOE is not making them open to re-use beyond existing ‘fair use’ policies in US copyright law. Under the DOE’s plans it isn’t clear whether the public can legally download ‘free’ manuscripts in bulk, redistribute them online, translate or content-mine their text, or otherwise mash up the text to create new works.
For some open-access advocates that is disappointing, because the OSTP directive called on agencies to “maximize the potential for…creative reuse to enhance value to all stakeholders”, says Peter Suber, director of the Harvard Office for Scholarly Communication in Cambridge, Massachusetts. That key phrase had even seemed to go beyond the NIH’s public-access policy since most papers at PubMed Central are not available to bulk download and text mine. Suber and others had hoped that the call for ‘creative re-use’ might lead federal agencies to announce open-licensing standards of the kind supported by the UK Wellcome Trust.
While the DOE doesn’t have legal authority to impose open-licensing standards on the publisher’s final version of a paper, it does have the right to redistribute DOE researchers’ accepted manuscripts for “federal purposes”, under what is called a Federal Purpose License. According to Suber, the DOE also has the right to authorize anyone else to do so.
But Hitson says: “DOE is not proposing that ‘federal purposes’ include giving the public broad authority to mass redistribute accepted manuscripts.” He adds that the White House directive instructed agencies to balance the services of publishers with the need for broad public access. “Rather than issue a blanket reuse license that may be subject to misinterpretation or even misuse, the [US] doctrine of ‘fair use’ gives the public substantial benefits in harmony with OSTP guidance,” he says.
Meanwhile, publishers say that the final version of papers made free on their websites will be available to read, download (individually) and analyse, just as the White House had required. The ability to text mine papers is also built into CHORUS, says Susan King, a member of its steering committee. SPARC’s Joseph says that she is nevertheless worried that “terms and conditions of use may be restricted” for articles on publishers’ websites.
Last week, the DOE also announced that from 1 October, researchers applying for funding will have to include a plan detailing how they will manage their data. All research data displayed in publications should be “open, machine-readable, and digitally accessible to the public at the time of publication”, the agency says; while “the underlying digital research data used to generate the displayed data should be made as accessible as possible to the public”. The other federal agencies have yet to make their plans public, but all have been submitted in draft form to the OSTP.