3.2 Repositories  (Page 2/3)

Rather than attempt an exhaustive analysis, it seems best to look at a number of categories of repository application: digital libraries, repositories of data sets, repositories of workflows and within workflows and repositories as Virtual Research Environments. Examination of each application will include concrete examples.

Repositories as digital libraries

While realising that the term is fuzzy, by digital library we understand here a system for managing, curating and delivering digital content that is primarily focused on (web) delivery to a human user, who is able to search, browse and access the material therein. While this also covers basic publication repositories, digital content may also be rich, varied and complex in structure, and digital libraries typically provide specialised access mechanisms and functionality for particular content types.

While some may dispute that the term should be so restricted – see for example the classic article ‘ What Is a Digital Library Anymore, Anyway? ’ – systems that provide functionality for manipulating, annotating etc. material may be better thought of as virtual research environments.

Examples of such repositories are many. Early examples are the University of Virginia Library or The Encyclopedia of Chicago , which use repositories to manage extensive collections of diverse material, including books, documents, images, maps and pre-existing websites. Many of these objects are themselves compound, and possess contextual relationships with other objects, all of which need to be represented in the data model. Different types of object are provided with particular access mechanisms that are required to be consistent across these object types.

More recently, there has been an emphasis on lightweight mechanisms for producing this sort of digital library. An interesting example is Active Fedora , which exploits RESTful web services (for an explanation, please see http://en.wikipedia.org/wiki/Representational_State_Transfer?) and rapid web development technologies to provide a low cost and low effort means for producing such applications, for example the Jewish Women’s Archive .

While a user may see only a website, there are important differences between using a repository and just using, say, a Content Management System (CMS); the structure of the information is entirely managed within the repository, and the services that support the behaviours and functionality of the information are associated with the repository objects. Both are decoupled from the online delivery mechanism, thus “future proofing” the content. Moreover, the repository provides a generic way of representing content, in contrast to specialised (often commercial) systems for delivering particular types of content, such as book digitisations, which provide high levels of specific functionality but less flexibility.

Repositories of datasets

Take up of repositories for managing more complex material occurred first in the arts, humanities and cultural heritage fields. In the sciences, the focus for data management was more on the need to manage, transfer and access large data sets distributed over multiple, collaborating research centres. The eCrystals archive, developed by the JISC-funded eBank project, was a pioneer in the use of digital repository technology – in this case EPrints – to archive, curate and disseminate research data, specifically chemistry datasets in the particular field of small-molecule crystallography. The repository here is integrated within the researcher’s work activities – datasets are deposited once the experiment is complete – and it exploits the publication mechanisms established for more traditional institutional repositories.