Internet has led to a fundamental increase of information that is avail- able to its users over the latest years. The users want to express their needs by simple means, such as queries and they want their queries to be evaluated without caring where the data are placed or how the queries are optimized. The work presented in this thesis contributes to the goal of declarative and efficient management of Web content in distributed settings and it is divided into two main chapters. In the first chapter we study OptimAX, an optimizer for the Active XML language which is able to rewrite a given Active XML document to an equivalent document which would, very likely, have smaller execution cost. With OptimAX we focus on the problem of distributed query optimization in the Active XML setting and we present two interesting case studies inspired by the R&D projects in which our group has been involved. In the second chapter, we propose solutions to the optimization problem from a different perspective. We optimize queries using a set of precomputed queries (materialized views). We have developed a peer-to-peer platform, called ViP2P (views in peer-to-peer) that gives to the users the opportunity to publish their XML documents and to specify views over these documents using a tree pattern language. Whenever a user asks a query, the system will try to find views that can be combined in order to find a rewriting equivalent to the asked query. We have carried WAN experiments that show the scalability of the ViP2P platform.
