Think Events: From Signals to Events

With more challenging problems arising in recent times, it became essential for data management research to start considering dynamic situations, in particular data streams [Babcock et al. 2002] and the events happening within them. Luckham and Frasca [1998] championed the concept of complex event processing in data streams, an idea that was adopted by many researchers and remains popular in traditional applications that process a few well-structured data streams for making real-time decisions. More challenging problems have pushed the concept of events to something that resembles more of what we see in foundational sciences such as in philosophy and linguistics. The concept of events and applications that consider events as important entities is now an emerging trend. Westermann and Jain [2007] proposed a six-facet model to represent event structure, attributes, and causality. Xie et al. [2008] proposed a 5W1H (What, Who, Where, When, Why, How) repre­ sentation to capture event attributes. In databases, Gatziu and Dittrich [1994] and Gehani et al. [1992] proposed models based on different event attributes. In most of these models (save for Westermann and Jain [2007]), causality and structure were not captured. The term event has been used in two distinct contexts in the computing litera­ ture: physical world occurrences and the representations of those occurrences in a computer system. In different computer science domains, event-based analysis is about capturing, processing, and managing low-level events such as publish/sub­ scribe systems and middleware solutions [Oberle 2006], complex event processing [Ericsson and Berndtsson 2007], event stream processing [Cetintemel 2003], and Think Events: From Signals to Events