Publishers, universities, and NIH vie for public access role

Publishers of scientific journals are proposing themselves as the vehicle for providing free public access to articles that result from publicly funded research. In a proposal made public on 4 June, publishers of more than 90% of the world’s scientific literature proposed a new database system they call CHORUS (Clearinghouse for the Open Research of the United States) that could be used by NSF, NASA, the Department of Energy and other federal research agencies to make the journal articles that result from their research grants freely available after an unspecified embargo period.

The CHORUS proposal is one of several options being pitched to the agencies, which face an August deadline for developing individual plans to implement a White House Office of Science and Technology Policy (OSTP) directive for free public access. The National Institutes of Health’s National Center for Biotechnology Information (NCBI), whose PubMed Central (PMC) database has for several years been the public repository for articles resulting from NIH-sponsored research, is proposing to expand PMC to encompass the articles from research paid for by other federal agencies. A third proposal being floated by universities and research libraries would provide public access by linking the hundreds of individual institutional digital repositories.

The CHORUS system would build on the CrossRef database, the existing platform for access to the content of more than 27 000 journal titles. A recently completed CrossRef pilot project called FundRef —a collaboration of seven publishers plus NSF, DOE, NASA, and the UK’s Wellcome Trust—uses publisher-provided metadata to automatically identify journal articles that come out of publicly funded research. Fred Dylla, CEO of the American Institute of Physics, one of the FundRef participants (and publisher of Physics Today), says that FundRef will become the essential back-end infrastructure for CHORUS. With the FundRef and CrossRef protocols, 80% of the infrastructure for CHORUS is already in place. Dylla estimates that building the user interface for the system and remaining components of CHORUS will cost about $1 million, to be funded by publishers who are members of CrossRef.

Saving money

Publishers say CHORUS would save the government money, but they acknowledge a self-interest in proposing to manage the public database. ‘We want the users to come to us for that content. It’s just the culture of the Web; we’d rather that they were on our platform than somebody else’s platform,’ says Patrick Kelly, director of publishing at John Wiley & Sons. ‘It’s the old eyeballs thing.’ Dylla says, ‘The whole goal of CHORUS is to take advantage of the infrastructure that publishers have in place to identify, provide access to, search, and archive, so agencies won’t have to invent and spend money on it. We’d much prefer that they spend their funds on research.’

John Vaughn, executive vice president of the Association of American Universities, one of three organizations that are proposing the Shared Access Research Ecosystem (SHARE), says that a potential collaboration is being discussed with the CHORUS publishers. SHARE would provide search and text mining of university research repositories plus lecture notes, conference proceedings and other ‘gray literature,’ he notes. Vaughn acknowledges that the CHORUS proposal is further developed than SHARE, whose other members are the Association of Public and Land-grant Universities and the Association of Research Libraries.

Officials at NCBI declined to comment on their proposal in advance of submissions of agencies’ plans to OSTP.

Yet to be determined is the duration of the embargo period during which publishers will have exclusive ownership of the publicly funded content. The OSTP policy directive was ambiguous; it set a one-year embargo as a guideline but added that agencies may establish a different period if they can justify it. Publishers have argued that a longer embargo is needed for less well-funded fields and for articles contained in infrequently published journals, so that librarians will maintain their subscriptions. Kelly says a two- to three-year embargo period may be appropriate for some social sciences and mathematics journals. ‘It’s important that we come up with some method of determining what the embargo period should be. Some sort of algorithm or metric based largely on usage so that some piece of an article’s usage, life, or half-life would drive the embargo period.’

‘The whole process of providing access to publicly funded articles means that the embargo period is long enough that you don’t hurt or jeopardize the subscription model, since that is how 90% of publications are still funded,’ says Dylla.