The HyperView Approach to the Integration of Semistructured Data
Lukas C. Faulstich · Universitätsbibliothek der FU Berlin Hochschulschriftenstelle u. Dokumentenserver · 2000
Using the World Wide Web to answer a specific question often requires information to be collected from multiple heterogeneous Web sites. Virtual Web sites are a promising approach to automate this task for particular, focused application domains. A virtual Web site serves pages containing concentrated information that has been extracted, homogenized, and combined from several underlying Web sites. The HyperView approach to the integration of semistructured data presented in this thesis provides a methodology, a formal framework, and a software environment for building such virtual Web sites. The HyperView approach treats the three steps of data extraction, integration, and presentation uniformly as consecutive views that map between different levels of abstraction. These levels are reflected by the architectural layers of the system. The contents of Web sites as well as the consecutive views are represented as graphs. Views are defined by sets of graph transformation rules. A demand-driven rule activation mechanism has been formally described and implemented. This mechanism incrementally materializes views in response to queries issued against them. The HyperView System has been implemented in Prolog. Graph transformation rules are compiled into efficient Prolog predicates. Java servlets are used to support virtual Web sites. The main contributions of this thesis are: 1\. the key idea of applying the same view mechanism uniformly to the problems of extraction, integration, and presentation, 2\. the HyperView methodology for modeling and integrating Web sites, 3\. the formal framework defining the data model, rule concept, and the demand-driven view materialization mechanism of HyperView, 4\. the HyperView System prototype providing a platform for building virtual integrated Web sites 5\. the validation of the HyperView methodology and system in case studies on Digital Libraries and Town Information.