How end-user development will save composition technologies from their continuing failures

Abstract

In this paper I discuss the motivation behind the continuing failure of composition technologies (workflows, service composition, mashups)over the past two decades, as well as the repeated mistakes we (the IT community, including companies) keep making. I then argue that, despite this, we do need these technologies, now more than ever, and that EUD is the driver behind this need, so the non-programmer has to be the target user for these platforms. Next, I share ideas as well as research results on how to design and develop composition tools that can be "programmed" by end user to define fairly complex logic both in the case of process-oriented applications and of mashups. These ideas are instantiated in a set of principles as well as in composition languages and tools developed in cooperation with our research and industrial partners. © 2011 Springer-Verlag.