Capturing Trust in Social Web Applications

Abstract

Abstract The Social Web constitutes a shift in information flow from the traditional Web. Previously, content was provided by the owners of a website, for consump-tion by the end-user. Nowadays, these web sites are being replaced by Social Web applications which are frameworks for the publication of user-provided content. Traditionally, Web content could be ‘trusted’ to some extent based on the site it originated from. Algorithms such as Googl e’s PageRank were (and still are) used to compute the importance of a website, based on analysis of underlying link topology. In the Social Web, analysis of link topology merely tells us about the importance of the informationframeworkwhich hosts the content.Consumers of information still need to know about the importance/re liability of the content they are reading, and therefore about the reliability of theproducers of that content. Research into trust and reputation of theproducers of information in the Social Web is still very much in its infancy. Every day, people are forced to make trusting decisions about strangers on the Web based on a very limited amount of information. For example, purchasing a product from an eBay seller with a ‘reputation’ of 99%, downloading a file from a peer-to-peer application such as Bit-Torrent, or allowing Amazon.com tell you what products you will like. Even something as simple as reading comments on a Web-blog requires the consumer to make a trusting decision about the quality of that information. In all of these example cases, and indeed throughout the Social Web, there is a pressing demand for increased information upon which we can make trusting decisions. This chapter examines the diversity ofsources from which trust information can be harnessed within Social Web applications and discusses a high level classification of those sources. Three different techniques for harnessing and using trust from a range of sources are presented. These techniques are deployed in two sample Social Web applications—a recommender system and an online auction. In all cases, it is shown that harnessing an increased amount of information upon which to make trust decisions greatly enha nces the user experience with the Social Web application