Overview
Executive Summary Providing open access to research output maximises research access which, in turn, maximises and accelerates research impact. The outcome is better research productivity for the individual research progress for the scientific community and hence increased rewards for the individual scientist and his or her institution.
Why Archive Research? In a research environment the imperative is dissemination - science itself is advanced by access to new knowledge and individual careers are judged by the impact that a person's contribution has on the rest of the community. The effect of the Web has been to increase the opportunities for access, allowing copies of research papers to be efficiently downloaded across the Internet. This improvement in access directly leads to an improvement in impact - as the readership of an article is increased then the probability that its work will be used and cited is increased. (Indeed, recent investigations confirm that papers that are available on the Web are more highly cited. Nature, Volume 411, Number 6837, p. 521, 2001.
Why EPrints? As the World-Wide Web grew in popularity in the mid-1990's, a number of sites grew up dedicated to disseminating academic research. Foremost among these was the High Energy Physics archive (at the time known as the Los Alamos archive after its then location at the US National Research Laboratory in New Mexico). At a meeting of interested researchers in October 1998, the issue of interoperation between these various sites was discussed and the Open Archives Initiative (OAI) began. At that meeting, EPrints was proposed as an off-the-shelf package that researchers could use to set up their own archives.
Why Not Just Use an Ordinary Web Page? One approach to archiving research is to upload it to your own Web page. If you take the trouble to make it available in an accessible browsing format (HTML or PDF rather than PostScript print job or a wordprocessor file) then people may eventually be able to find the document through Google and read it directly in their Web browsers. The drawback with this approach is that the information that people use to find the article may not be available. For example, a Physicist may be trying to find Smith H. J. Phys A: Maths Gen, 27:1425, 1996 and have no way of relating it to the URL of the article on the author's web site. Physics is particularly vexing in this respect because bibliographies do not use articles' titles. Consequently, even Google will not help to locate this paper!
The reason for the use of EPrints software rather than a simple Web server is the need to capture the publication information about each article - its authors, title, date of publication, journal, issue and page numbers. This metadata is provided by the author when the paper is initially uploaded and stored with the paper. Subsequently his metadata is harvested by other OAI services (search engines, citation analysis agents etc.) so that the collected information from individual archives can be used to create 'literature-wide' services.
Why Self-Archive? The significant issue with archiving is that it is the creator of the work that makes it available. If you take someone else's work and put it on the Web, that is Napster-style theft. If you make your own work available then that is dissemination. You may delegate the process of uploading the article and filling in the metadata to a student, a colleague or an administrative assistant, but ultimately it is the author who is responsible for self-archiving.
Glossary
An eprint is the digital text of a peer-reviewed research article, either before or after refereeing and publication.
A preprint is any version of an article before the final, refereed, revised, accepted draft.
A postprint is any version of an article from the refereed, accepted, final draft onwards (including post-publication corrections and revisions).
Metadata is information about an eprint, usually the name of the authors, the title, date, journal etc.
To self-archive is to deposit a digital document you have written in a publicly accessible website. EPrints is an OAI-compliant web server which provides a simple interface for the depositer to copy/paste the important metadata for an article as well as attaching the full-text document.