The DOI scheme needs to be extended so that each DOI can refer both to the article hosted at the journal and at the institutional repository. Each university needs to have a mechanism to look up an article DOI against the open access articles held in their repository.
We need some uniform way of resolving every document address
(e.g.) DOI:10.1103/PhysRevLett.89.117901
into the publisher’s (paywalled) and the researcher’s (free to view) copy.
The host institution needs to respect the 12 month embargo, e.g. only host the abstract for the first 12 months, so that the business model of journals isn’t completely cannibalised.
I think the public is reasonably well served by what’s effectively a 12 month monopoly on distribution by held by the journal, e.g. the public has an interest in ensuring that good science takes place in future, as well as in finding out the results of science that’s already been funded. It’s perhaps right that journals have an economic incentive to pick papers that are important enough that people will pay to read within the first 12 months of being published, because that encourages them to select good science.
The long-tail business model of selling the back catalog of possibly decades old publicly funded research involves almost no effort, judgement or resources from the publishers once an article is digitised, so it should be discouraged.
We also need a scheme whereby any journal article cited by an open access article is brought out from behind a paywall. The references to an article need to be available to make use of and understand the original.
One possibility would be to have a government grant to publishers in the same way that the scientists receive government funding. The journals do useful if slightly mundane work in selecting good science, in rejecting flawed studies, and in moderating peer review - it’s a function that needs to be continued in some manner for science to be conducted, irrespective of how the execution of that function is paid for.
If the government sponsored journals to host open access at the rate of 1¢ per unique download per day, and say $100 for each time an article is cited by another published work, together with a $1-$3K fee for handling the manuscript, you could create an economic incentive for journals to select the best work to publish, and still ensure that studies become available to the people (taxpayers) who’ve paid for the research.
