Design, Implementation and Validation of a Middleware Enabling the Exploitation of the Web of Documents

This thesis presents and details InterDataNet, a framework for information aggregation in the form of graphs of individually addressable information pieces. The framework pursues the objective of creating a read/write Web of Documents, by exposing these graphs as resources called IDN-Documents whose vertexes can be manipulated with CRUD operations. Graphs can be linked together and new IDN-Documents can be built by reusing parts of different IDN-Documents. Moreover, several properties are enabled both at IDN-Document and vertex level. The framework is made up of an information model, called InterDataNet Information Model, which regulates the representation of InterDataNet entities such as IDN-Documents, IDN-Nodes and different types of Links (Aggregation Link, Reference Link, etc.). Such information model is implemented through the InterDataNet middleware, a fully RESTful layered architecture providing services for the management of IDN-Documents, names, history and persistence. A detached module, the Adapter, provides adaptation capabilities towards external data sources to represent outer data as IDN-Documents. InterDataNet is also provided with the Activity Node, a scriptable vertex for dynamically generating data, using other vertexes' contents as inputs. Finally, the framework includes a tool-suite, to support developers in exploiting framework capabilities. The tool-suite consists of the IDN Java and IDN.js libraries for the server-side and client-side exploitation of the IDN-Document, respectively, a visual document editor called IDN-Studio, and IDN-Viewer a JQuery plugin for rendering IDN-Documents as HTML pages. InterDataNet has been selected to participate in the SmartSantander FP7 EU project, to provide Web of Resources capabilities in a real-word Smart City scenario. To this end, sensor data coming from the experiment platform has been represented as IDN-Documents, and added-value applications have been built on top of them. This experiment has proven useful for InterDatNet validation, and qualitative and quantitative evaluations have been included.