A Web Mashup for Social Libraries Domenico Gendarmi, Filippo Lanubile University of Bari Dipartimento di Informatica Via E. Orabona, 4 - 70125 Bari +390805442286 {gendarmi, lanubile}@di.uniba.it ABSTRACT User-generated content on the Social Web is often locked within information silos. Inadequate APIs or, worst, the lack of APIs obstruct reuse and prevent the opportunity to integrate similar content from different communities. In this paper we present a Web mashup which combines information from different social libraries. Aggregated information, including both classic book metadata and user-generated content, is represented as linked data in RDF to allow machine computation and foster reuse among different applications. Categories and Subject Descriptors H.5.3 [Information Interfaces and Presentation]: Group and Organization Interfaces – Web-based interaction. General Terms Design, Human Factors. Keywords social libraries, mashup, linked data, social web, semantic web. 1. INTRODUCTION The rise of the Social Web, also known as Web 2.0 [8], created new challenges and opportunities for user interface integration. In such a context, Web mashup has emerged as a dynamic approach for composing content and functionalities originating from disparate web sources [10]. Nowadays, plenty of user-generated content is available on the Social Web. However, these forms of collaborative contributions are restricted to one single application and do not let data flow from one point on the Web to another. Furthermore, most applications’ creators on the Social Web are quite reluctant in providing programmatic access to user generated content which is hosted within their web sites. As a consequence, often no APIs are made available. Although some programming interfaces are sometimes provided, third-party applications are only allowed to retrieve information concerning one user at time, thus hampering the reuse of community-generated content and the integration of similar content from different communities. Social libraries, also referred to as “Libraries 2.0” [7], can be regarded as an innovative form of library services which replace traditional ones. Social libraries attempt to harness users in the design and implementation of the application by encouraging feedback and participation. As part of the Social Web, these applications are also affected by inadequate APIs or, worst, the lack of APIs. We thus paradoxically face a situation where one book can be tagged and reviewed in Amazon, LibraryThing, and Anobii, but there is no way to know that people refers to the same paperback. In this paper, we illustrate a Web mashup, named Colibrary, which combines classic book metadata (e.g., title, authors, publisher) and user-generated information (e.g., tags and reviews) from different social libraries. The main goal of the mashup is to combine such data from different communities of book readers, as they can be shared rather than locked within ‘information silos’. Shared data are structured as linked data, thus represented through semantic web technologies, in a manner that allows machine computation and fosters reuse among different applications. 2. COLIBRARY MASHUP Colibrary initially originated as a Web API to exploit the opportunity to reuse content in the digital library domain and provide other applications with different types of information about books. Since its inception, the basic requirement was the possibility to unambiguously distinguish one single paperback. A straightforward solution was the adoption of the International Standard Book Number (ISBN) as a unique identifier for referring to a specific book (Figure 1). Figure 1. Colibrary Overview Colibrary then evolved as a server-side mashup [6]. Once a valid ISBN is given as input to Colibrary, data about the specific paperback are retrieved from three large web sites where different community of readers can organize their virtual bookshelves by sharing keywords and reviews. In particular, Colibrary collects the following bibliographic data for a book: title, author(s), editor(s), publication date, number of pages and the cover, whether available through the Amazon Web API. Then, Colibrary supplements such information with social data directly generated from end users of popular applications in the library domain. Specifically, Colibrary retrieves all the tags and reviews assigned Copyright is held by the author/owner(s). WWW 2009, April 20-24, 2009, Madrid, Spain.