Writing has always been the storehouse of human knowledge and the backbone of civilisations. To this day, it carries our best hopes for building a better future. Within it is the potential to curb climate change, stave off drought and famine and cure diseases.

However, the flow of information is stifled by the very means we use to disseminate it: publishing. Scientific publishers wield enormous influence over society through the control they have over the direction of science. That control comes from their power to make or break scientists¡¯ careers; the primary goal of much research today is not the advancement of knowledge or society but rather getting papers into the right journals.

True, getting published does typically require the discovery of something novel. However, social and technological innovation occurs at a snail¡¯s pace because progress has been rendered a side-effect of research, rather than the goal. One immediate consequence is the pressure researchers feel to prioritise individual achievements and monopolise their data to maintain a competitive edge for future publications and funding applications.

We propose a new vision for scientific publishing that starts with reversing the relationship between authors and publishers. Under this system, authors would be able to make their research freely accessible to everyone immediately. Journal editors would compete to publish it, but publication would not be the end of the story: researchers could continue to update their papers for years afterwards. Nor would publication be the aim of the game: the incentives, recognition and reward systems would not depend on where a paper is published, but rather on its contents and the extent to which it advances knowledge.

³ÉÈËVRÊÓƵ

This is already starting to happen. The number of preprints is increasing daily, and most journals now facilitate the submission of papers to preprint servers via their own submission systems. Others have appointed to screen preprints and solicit submissions, adopting scoop protection policies that commit them to disregarding, in their editorial decisions, any competing papers published after submission of the paper or preprint.

But the evolution needs to go further, faster. All papers should be indexed and retrievable via a single open database, clearing the way for feedback from students, citizen scientists and researchers at all career grades and compass points. This process would allow authors to address any obvious concerns before journal publication, updating or correcting their papers with comments, new data or links to improved methods or protocols.

³ÉÈËVRÊÓƵ

This change is all the more necessary because papers often contain specialised data from diverse disciplines, many of which are beyond peer reviewers¡¯ expertise. This means that the papers are not thoroughly reviewed at the experimental level.

Eventually, we hope that this process will obviate the need for traditional peer review. But we recognise that experiments with post-publication peer review have had limited success so far and that it will take time to instil new incentives and a new academic culture in which commenting on preprints is the norm ¨C and is trusted to enforce standards and editorial remits. But, at the very least, all reviews and authors¡¯ responses should be published alongside the paper.

To be fair, some journals have already adopted this practice, too. Others require the deposit of supplementary data (although rarely of the tools that would allow others to validate and expand on the work). This should become standard practice ¨C as should the writing of brief, genuinely lay summaries that explain the motivation of the work, the main results, their limitations and their implications for human knowledge and/or society.

Our proposed system would eliminate the time lost on multiple submissions and rejections, as well as delays resulting from reviewers' demands for more experiments: particularly important for junior scientists, who need to build a reputation quickly.

³ÉÈËVRÊÓƵ

Scientists are currently reluctant to invest time in correcting or improving published work, by either themselves or others, because it brings little credit. Indeed, new information has to be submitted via a correction, which has negative connotations. A dynamic publication system that keeps track of all these contributions is a crucial prerequisite for a culture that rewards them. It would enable scientists to regularly update their publication with improved protocols and data that are important in relation to the original discovery but cannot be assembled into a complete story. Such research is currently not made available because there is no reward in doing so.

Plos Biology and Plos Computational Biology that 39?per cent and 46?per cent of their authors, respectively, opt to make preprints available ¨C and that many who do not are simply unfamiliar with the process. That unfamiliarity could soon disappear as funders pick up the baton. Some ¨C including the European Research Council, the Wellcome Trust and the National Institutes of Health ¨C already encourage scientists to cite preprints in grant applications.

But fully establishing this brave new world will require further concerted efforts by all stakeholders. More investment is needed in developing the required publishing technologies, for instance, and ensuring the long-term sustainability of repositories (private funders could play a big role: witness the to BioRxiv by the Chan Zuckerberg Initiative).

Most crucially, though, universities and research institutions must begin experimenting with new evaluation systems that incentivise scientists to embrace the advantages offered by the new publishing system. Otherwise, researchers will continue to prioritise publications over people and society.

³ÉÈËVRÊÓƵ

Hilal A. Lashuel is associate professor of life sciences at ?cole Polytechnique F¨¦d¨¦rale de Lausanne (EPFL). Benjamin Stecher is an independent patient advocate diagnosed with Parkinson¡¯s disease.