When it comes to science, the coin of the realm, the measure by which ideas and scientists are evaluated, is by being published in peer-reviewed journals. The journals are run either by professional organizations of scientists as a non-profit arm, such as all the journals run by the American Institute of Physics or the American Physical Society or are purely commercial private operations, such as the prestigious journal Nature. Journals subscriptions are then sold to individuals and libraries. But over time, a combination of rapidly rising subscription costs, shrinking library budgets, and the rise of the internet has created both a crisis and an opportunity for radical changes.

A paper that appears in a high-quality peer-reviewed journal counts a lot both in terms of the credibility of the research and towards the career of the authors. But what exactly happens in this process? Before a paper is published in a reputable journal, the editors send the manuscript to other scientists familiar with the field to review the work and then, based on the feedback received, the editors decide whether to publish or not, or publish after revisions. In order to encourage frank reviews and reduce the effect of biases, this process of peer review is conducted anonymously in that the author of the paper does not know the identity of the reviewers and, less commonly, the author’s identity and institutional affiliation are stripped from the paper before it is sent to the reviewers, so that only the editor knows the identities of the author and reviewers.

But it is important to realize that a decision to publish does not mean that the reviewers and editors are certifying that the work is necessarily correct. Reviewers are not required to replicate the work and confirm the results. What they are expected to examine is whether the work says anything new and worthwhile, conforms to the standards in the field, properly cites other work, follows standard protocols, is internally consistent, has no obvious errors, makes sense, is essentially sound, takes into account implications for related work, and lays out all the information that would be necessary if someone else were to try and replicate the work.

What is important to realize is that the people actually engaged in the heavy scientific work (the authors, the editors, and the reviewers) all volunteer their time and energy. No one gets paid because they usually have research jobs, though some editors might get a small stipend. Indeed some journals expect the authors to pay a page fee to offset some of the cost. (My relatives were baffled that I would go to all the trouble of publishing a paper without getting paid for my efforts.) The value added by the journal was to act as the clearing house for the papers to go between authors, editors, and journals, and, in the old days, to take the printed manuscript and typeset it for printing in the journal.

With the advent of digital publishing, the main work of journals of taking a typed manuscript and converting it to print form has disappeared. Now authors send in camera-ready documents and the journals themselves are available in electronic form. Since pretty much everything is done electronically and for free, it has become less clear why we need to pay for journal subscriptions, which have become exorbitant. This has led to the rise of open-source journals and repositories of preprints, bypassing journals altogether. Some subscription journals have tried to respond by putting up a pay wall for a limited time, after which the article becomes open access.

Now comes a report that a consortium of major research funders in Europe are demanding that their grant recipients only publish in open-access journals.

Research funders from France, the United Kingdom, the Netherlands and eight other European nations have unveiled a radical open-access initiative that could change the face of science publishing in two years — and which has instantly provoked protest from publishers.

The 11 agencies, who together spend €7.6 billion (US$8.8 billion) in research grants annually, say they will mandate that, from 2020, the scientists they fund must make resulting papers free to read immediately on publication (see ‘Plan S players’). The papers would have a liberal publishing licence that would allow anyone else to download, translate or otherwise reuse the work. “No science should be locked behind paywalls!” says a preamble document that accompanies the pledge, called Plan S, released on 4 September.

“It is a very powerful declaration. It will be contentious and stir up strong feelings,” says Stephen Curry, a structural biologist and open-access advocate at Imperial College London. The policy, he says, appears to mark a “significant shift” in the open-access publishing movement, which has seen slow progress in its bid to make scientific literature freely available online.

As written, Plan S would bar researchers from publishing in 85% of journals, including influential titles such as Nature and Science. According to a December 2017 analysis, only around 15% of journals publish work immediately as open access (see ‘Publishing models’) — financed by charging per-article fees to authors or their funders, negotiating general open-publishing contracts with funders, or through other means. More than one-third of journals still publish papers behind a paywall, and typically permit online release of free-to-read versions only after a delay of at least six months — in compliance with the policies of influential funders such as the US National Institutes of Health (NIH).

Subscription journals argue that they are necessary to maintain quality control but that argument is becoming less credible as alternative ways of evaluating research quality are being developed. Ultimately, the test of scientific ideas is whether those ideas work and are useful and we should not need high-subscription publications to ensure that.