Materialization of web data sources

Recent years witnessed an exponential increase in the number of data services available on the Web. Many popular Web sites, including social networks, offer API for interacting with their information, and open data initiative such as the Linked Data project promise to achieve the vision of the Web of data. Unfortunately, access to Web data is typically limited by the constraints imposed by the query interface, and by technical limitations such as the network latency, or the number and frequency of allowed daily service invocations. Moreover, several sources may independently publish data about the same real-world objects; in such case, their combined use for assembling all available information about those objects requires duplicate removal, reconciliation and integration. This paper describes various data materialization problems, defining properties such as source coverage and data alignment of the materialized data, and then focuses on a specific problem, the reseeding of data access methods by using available information from previous calls in order to build a materialization of maximum size.