Article Preview
TopIntroduction
Web applications (“Web apps”) have been developed very rapidly over the past years. Their development is different from the development of “classic” applications in that Web apps are perceived and delivered as services, which “impacts the entire software development and delivery process” (Musser & O’Reilly, 2007). What is imminent to both kinds of applications, traditional and Web, is that there can never be a full accordance between providers, users, or other stakeholders as to which features a certain application should have. The utility of new features and functions differs between the stakeholders, and there will always be functions that will not be implemented by the provider (although they are requested by users). With their ever growing importance, the need for customizability henceforth also increases. Moreover, as Macías and Paternò (2008) note, “there is an ongoing shift to end-user centered technology, and even users with poor or no skill in Web-based languages may feel the need to customize Web applications according to their preferences.” The present paper studies an approach to Web application customization that fits particularly well with frameworks whose purpose is to integrate mashups into Web apps.
The problem of application customization is particularly interesting for the context of Web applications. Karger et al. (2009), Lutteroth and Weber (2008), and Macías and Paternò (2008) each deal with different aspects of this. A key observation of Karger et al. (2009) is that “instead of warping their data to fit rigid applications, users should warp applications to fit their data and tasks.” The present paper takes a complimentary approach by suggesting practical Web app enhancement through mashups. In the first part of this paper, we present the ActiveTags prototype originally outlined in Hagemann and Vossen (2009), which enables this based on tag-based mashups that are integrated into Web pages on the fly via a browser extension. In the second part, the generality of the approach is documented by a transformation of the application into a relational equivalent.
Attempts to cope with the challenge of application customization from a technical point of view are Application Programming Interfaces (APIs), which allow skilled developers or users to create custom functionalities based on the data and functions provided by Web applications. The number of such API has been rising continuously over the last years. This allows for what Hippel (2005) has called “democratizing innovation.” Well-designed API turn Web sites into platforms that make themselves indispensable by providing the grounds for “ecosystems” of applications (cf. Musser & O’Reilly, 2007).
One form of API usage is mashup creation, i.e., Web applications which combine data or functions from one or more sources into one interface. Because of the perceived tentative character, mashups were first considered to be merely toys or gimmicks, but gradually they are receiving more recognition from both research and industry (cf. Ennals et al., 2007). The utilization of mashups in enterprises has been hampered by some important barriers which prevent widespread adoption (cf. Hinchcliffe, 2007). However, many companies, large and small, have started to turn to mashups as a useful option to quickly attain some needed functionality. In fact, the financial crisis that took off in late 2008 has strengthened the role of mashups in companies seeking cost savings and improved operations (cf. Rodier, 2009). Offering API enables third parties to develop functionality externally. However, the prerequisites are still high, as they require programming skills, knowledge of the particular API, and the appropriate infrastructure to run on.