How To Achieve Open Access (Part 1)

This is a description of how to improve scholarly communications by providing toll-free access to peer-reviewed journal (and conference) papers. See the preamble for a more detailed discussion of the target and particularly for those things which are not the target of this guide.

Papers can be deposited in a Repository: a database designed to be accessible online which includes a minimum set of meta-data1 describing a paper (author names, title, journal or conference title, year of publication). There are good reasons for universities and similar research institutions to run their own repositories: an institutional repository or IR. The small number of other authors can register for free with the Open Depot. Having a repository is only half (perhaps less than half) of the solution, however, since most authors for a variety of reasons will not spontaneously deposit their papers. The vast majority of these papers are written by people working for universities and similar research institutions. In order for near-universal access to be achieved, therefore, such institutions need a strong policy to require staff to deposit their papers. This needs promotion by knowledgeable and passionate staff (preferably a mix of librarians and researchers) as well as coherence with other institution policies:

  1. Immediate Full Text and Minimal Meta-data Deposit of the Author's Submitted Text

    All staff who publish reviewed papers in academic journals (or conference proceedings for disciplines like CS with reviewed full-paper conference proceedings) must deposit their own submitted text plus minimal meta-data (title, author name, journal or conference name, year) in the institutional repository on acceptance.

    The policy must not try to require the publisher's formatted PDF or any copyediting, just the author's final text - authors can include any copyedits or broader meta-data if they wish but the policy must not require it.

  2. Immediate Deposit/Optional Access

    Having deposited the paper and minimal meta-data in the IR, where the publisher allows (see SHERPA ROMEO for a list) the full text should be set to be accessible. Where the publisher requires an embargo or does not allow for open access, the full text must still be deposited but set to closed access in the repository software (both eprints and dspace have this option). Ensure that the repository has enabled the "request an eprint" functionality of the IR software and encourage authors to provide eprints whenever they receive an email request via the IR (the email the IR sends to the author includes a one-click way of using the IR to send the paper in reply).

  3. Coherence and Cross-reinforcement with Other Policies

    For tenure, promotion and any other staff evaluation processes for which lists of publications are required, revise the forms and processes to accept only papers which have been deposited in the IR. This requires limited staff time to check - require them to include the IR URL of each paper in the list. A little development work can provide a simple way of selecting the relevant ones from a particular author's papers in the IR, from which a standard format listing is produced for copying and pasting into the submitted evaluation form.


There are many side-benefits that a repository can produce such as automatically producing web pages listing papers from projects, from individuals, from departments, in disciplines etc. These, however, are only really about using the meta-data and a focus on these side-benefits can distract from the first and most important thing which is to require deposit always and open access where possible, plus simple manual provision where open access is not possible.

Do not require or even encourage authors to publish in open access journals just because they are open access. Encourage authors to publish in the right journal for their work and intended audience (including kudos for highly selective journals) and then require them to deposit their papers in the IR. Authors should submit to the most appropriate venue for their work and then make the text available in the IR.

Do not allow papers in open access journals to be not deposited (it makes checking compliance too complicated - if every paper must be in the IR then every paper must be in the IR - if only a proportion of papers must be in the IR then checking compliance requires lots more work than the deposit actually requires).

Do not attempt to change the copyright transfer process which most publishers require. It's not necessary for providing access to the text and while desirable in the long term introduces too much complication now and retards the delivery of the "within our grasp" access now. Better bread today than bread and jam tomorrow!

Do not put any librarian or other administrators in the way of the deposit for "quality checking", "publisher policy checking" or anything else. These things can be done if resources allow, but let the authors deposit, make that deposit visible (open access to the text or with the button request, as appropriate) and make it visible on the web (and particularly to Google Scholar and Microsoft Academic Search etc.). If you have the resources and want to do it, administrators/librarians can add additional meta-data etc, but don't allow bureaucratic limits to get in the way. Don't let a lawyer worry you with tales of copyright infringement lawsuits. No publisher has ever sued a university over making their academics' papers available. At most you need to respond to "take-down" requests by setting access to closed instead of open (never remove a paper from the IR, just set its access as closed, unless the paper is formally withdrawn by the journal for academic misconduct).

Once your institution has set up an IR and adopted this policy, what next?

1 Additional meta-data can be useful but is not necessary for providing access and lack of anything else should never get in the way of providing access to the full text with the minimal reference metadata.


Valid HTML 4.01 Transitional

Last modified: Tue Mar 10 20:07:06 JST 2015