A schema-based FavoriteWebsites service for centralized access to per-user favorite website data, wherein access to data is based on each user's identity. The FavoriteWebsites service includes a schema that defines rules and a structure for each user's data, and also includes methods that provide access to the data in a defined way. The FavoriteWebsites schema thus corresponds to a logical document containing the data for each user. The user manipulates (e.g., reads or writes) data in the logical document by data access requests through defined methods. The structure of the data is defined from the perspective of the data, not from that of an application program or a device, whereby appropriate programs can communicate with the FavoriteWebsites service to access the data, with existing knowledge of the schema-defined format, regardless of the device or application program in use.