The Wellcome Trust plans to withhold a portion of grant money from scientists who do not make the results of their work freely available to the public, in a move that will embolden supporters of the growing open access movement in science. In addition, any research papers that are not freely available will not be counted as part of a scientist's track record when Wellcome assesses any future applications for research funding.As one of the major donors to scientific research in the UK, this is bound to have an effect. I support this; science requires that knowledge is shared, so that advances can be checked and built upon by multiple researchers. Repeatability is at the heart of the scientific method; while the possibility that anyone can contribute ensures both the highest level of scientific development, and protects science against cultural biases and blindspots.
The problem is the position of scientific journals as the gatekeepers of the knowledge. Published work remains the gold standard for assessing a scientist's output. The review process serves to catch and correct mistakes and omissions, and to check that a research paper says something genuinely original. The journals in turn need funding to maintain themselves, so must charge a fee. The low print runs mean that the cost of a journal is quite high, and if the work in a journal is freely available elsewhere, there is little incentive to buy them.
In my field, it could be said that we work quite well without journals. After all, the first place to find papers in theoretical physics is the preprint arXiv. By the time a paper actually gets published several months later, there might be a number of responses to the original work. A big reason for this is that it is quite easy for a theorist to judge if a theoretical paper is reasonable or not. Having made our judgement, we can then decide if we want to respond or otherwise follow up on it. (With the volume of work produced daily, the answer is usually no :)
However, despite this most people still submit their work to a journal. And this is still useful; mistakes can be caught, areas where the work is unclear can be expanded on, omissions in the citations can be filled. Indeed, my last paper had a fairly embarrassing oversight that was caught by the referee, and the paper was definitely improved as a result. What's more, we can only judge the merits of a new paper because we are constantly immersed in the field; those who are not will not be able to make the same judgements. To turn things around, even in the very similar field of experimental particle physics, there are many areas where I cannot properly judge the merits of a paper as it is outside my expertise. The review process serves me in such cases.
So it seems that we cannot get away from some type of journals. But this returns us to our original problem: how can we ensure the freedom of knowledge and ensure journals remain sustainable? In my field, we seem to have stumbled upon a simple solution. Universities agree to buy subscriptions from journals, and journals agree not to complain about work being put up online. But it's difficult to see this happening in other fields. One possibility is for scientists to pay for putting their work in a journal; this may be possible in some fields, with larger grants, and indeed the Wellcome Trust is willing to provide for this in its grants. Another possibility is to make all work available after a short time, say six months. Universities would need to continue to buy subscriptions, so that their employees would have access to cutting edge research. But the interested citizen could still get access to the fruits of science.
So at the end of it all, where do I sit? I prefer the situation in my own field, for its simplicity. But if that is not possible, then grant agencies and Universities might need to realise that the payment for journals should come up front, at acceptance, rather than at publication. The modern world demands that published research be available to inform all people, for the good of both society and science.