Abstract

Since the Web 2.0 advent, Web users have gained more and more power moving their role from simple information consumers to producers. In line with this trend, mashup technologies have immediately attracted a lot of attention and many research investments since their birth. The big mashup promise was to bring application development to the masses, so that any Web-educated worker, also non-IT skilled, could implement his/her own situational applications (i.e., relatively small applications meant to address a temporary need of one or few persons) exploiting the simple paradigms and visual metaphors provided by mashup tools. After a decade from the first mashup tools, though, mashups are not really part of people’s everyday life and are still rather unknown technologies that — beside some exceptions — hardly find concrete application in the real world. During our research in this field our high-level goal was to foster the adoption of mashup technologies by end users. Aiming at this, we identified three main characteristics that must be reached by mashup technologies to get to the expected diffusion: (i) usefulness and (ii) usability for the end users and (iii) affordability for the developers of the respective mashup tools (in terms of required skills, time and cost). We identified lacks in these achievements as main hindering factors for the wide adoption of mashup technologies. Making mashup technologies useful, usable and affordable, therefore, are the three main challenges we addressed in our research work. This work contributes to the achievement of all these three major goals: first, by enabling a so-called universal integration paradigm, focussing on the creation of more powerful and complete mashups allowing data, application logic and user interface integration in one single language and tool; then, by introducing and developing domain specific mashup technologies, able to lower mashups’ complexity and make them usable by domain experts (i.e., end users expert of a given target domain); finally, by realizing a system able to generate domain specific mashup platforms as a service, basically relieving developers of platforms implementation and, therefore, making platform development affordable. This thesis specifically focusses on the last two points, i.e., on the domain specific mashup approach and on the semi-automatic generation of domain specific mashup platforms.