DSpace at MIT: What Users Want

This article was published in the October-November 2000 SPARC e-News, the online newsletter of the Scholarly Publishing and Academic Resources Coalition, a "worldwide alliance of research institutions, libraries and organizations that encourages competition in the scholarly communications market".

DSpace is a joint project by MIT Libraries and the Hewlett Packard Company to build a stable and sustainable long-term digital system that captures, preserves and communicates the intellectual output of MIT's faculty and researchers. The project provides an opportunity to explore issues surrounding access control, rights management, versioning, retrieval, community feedback, and flexible publishing capabilities. Launched in the Spring of this year, the project is expected to produce a functional system by the Fall of 2001. In the future, DSpace could serve as a model for other institutions, resulting in a federation of systems that make available the collective intellectual resources of the world's leading research institutions.

Existing e-print archives have been organized around specific disciplines, for instance, the LANL ArXiv for physics and mathematics. Researchers from many institutions contribute to these archives. DSpace, however, is an attempt to provide a means of e-print dissemination using a multidisciplinary institutional model. The DSpace system will address the issue of disciplinarity by subdividing DSpace into "subspaces", each of which will serve a particular discipline or type of information. Subspaces will be managed by the members of the subspace community. Each subspace will develop its own policies concerning who has the right to contribute content, how or if submissions will be reviewed, what metadata will be required, and who can have access to the subspace content.

The first stage of this two-year project has included an ongoing fact-finding and needs assessment process consisting of extensive interviews with prospective content-contributors to DSpace. The goals at this stage are:

• To find out who will contribute content to DSpace
• To find out what types of documents will be contributed
• To find out how subspaces can be managed
• To learn about workflow in the process of submitting papers for publication
• To find out how users prefer to interact with DSpace
• To create awareness of DSpace at MIT
• To foster buy-in and cooperation
• To identify early adopters and willing usability study participants

So far we have held interviews with faculty in the Departments of Chemistry, Brain & Cognitive Sciences, Mathematics, Ocean Engineering, Chemical Engineering, Mechanical Engineering, Civil Engineering, Physics, Linguistics, History, Music, Architecture and Foreign Languages and Literatures.

Faculty responses to the DSpace project have fallen into two distinct categories depending upon the publishing conventions and "cultures" within their disciplines. In disciplines where it is established practice to submit one's papers to electronic preprint archives, such as Physics and Mathematics, faculty expressed no concerns regarding peer review, nor were they concerned about jeopardizing their chances of publishing in established journals. These faculty roundly dismissed journals that reject articles based on previous appearance on a website. Some faculty claimed that the feedback they get from dissemination on the preprint server is more valuable than comments made by the limited number of reviewers provided by publishers. They claimed that preprint servers were more important than traditional journals. On the other hand, these faculty did not see a strong need for an MIT-based system, since they already have adequate facilities for preprint dissemination through their disciplines. We were assured, nevertheless, that many of them would also contribute to DSpace, as long as we made the submission process simple. Ideally, they would like to be able to submit to their discipline-oriented e-print sites and DSpace using exactly the same procedure, preferably in one step.

In most other disciplines, where there is not a strong e-print culture, faculty expressed a strong belief in the peer-reviewed publishing process, and also expressed strong concerns about the quality of a system such as DSpace. Many expressed concern about copyright transfer agreements they sign with publishers that preclude or limit the right to display articles on personal or institutional websites. Others mentioned that their need for timely dissemination of research results was met by publishers' "preprint" sites, where accepted-but-not-yet-published articles are displayed.

Faculty in all disciplines were interested in using DSpace to display images, datasets, video and audio files that they don't publish in established journals. Many expressed dismay at the mounting "page charges" required by publishers, especially for color graphics and images. As more and more research is being expressed in rich-media formats, DSpace will offer faculty a means of publishing an unlimited amount of digital information in formats that are not usually handled by traditional publishers.

Since a significant amount of MIT research is centered around Labs and Centers throughout the Institute, the DSpace needs assessment process included interviews with faculty and staff in the following research groups: the International Motor Vehicle Program, the Lean Aerospace Inititative, the Center for Real Estate, the Lab for Information Decision Systems, the Industrial Liaison Program, the Center for Energy and Environmental Policy Research, the Center for Global Change Science and the Center for Innovative Product Development.

Many of the labs and centers have already mounted publications on their websites, often with the motive of providing industrial sponsors with timely access to research information. Some of the labs have a history of publishing regular series of research reports and working papers; many have made the switch from print to electronic dissemination.

A common theme among these labs and centers is that one person is responsible for managing the electronic dissemination of publications, usually with some kind of sporadic technical help. Policies for content submission, access to publications and workflow procedures have already been established. Many of them expressed problems in maintaining a regularly updated electronic publications website, citing lack of sufficient server space, lack of technical support, lack of staff to update the site, and inability to provide adequate search capabilities. Almost all of these groups have expressed enthusiastic interest in DSpace as a potential means of solving these problems. They view DSpace as a reliable central service through which they can implement the e-publishing process. Because they have some idea of the cost of such a service, several centers have even expressed a willingness to pay for DSpace services.

The timeline for DSpace development includes a target date of early January for development of "Little D", the experimental prototype for DSpace. Concurrent with this effort, we are determining user preferences for system features using surveys of lead users. We also plan to test several user interface prototypes as the team works towards development of the MIT product. Updates on these efforts and further progress can be found at http://web.mit.edu/dspace.

Margret Branschofsky Faculty Liaison, DSpace Project, MIT Libraries margretb@mit.edu

Contact Webmaster