Mark Hapner, Distinguished Engineer – SOA Strategist, Sun Microsystems
To summarize, Mark says, “Long before the Internet, event based application integration was done with proprietary Message Oriented Middleware. Now we are building global applications but we don’t have a practical global replacement for MOM’s event infrastructure. This talk describes how RSS, and in particular, the latest form of RSS – Atom – is the event infrastructure for global applications.”
In my opinion, the CEP RSS/Atom use case proposed by Mark is one of the more interesting event processing use cases we have seen to date. In addition, Progress Apama and the good team of Mark Palmer, Dr. John Bates, Andy Menzies, and Dr. Giles Nelson, et al have already started down the path of news-event processing as a commercial CEP application with their global elementizing deal with Dow Jones, announced a few months ago.
The good news is that is should, in theory, be easier to process RSS/Atom events that human speech perception and meaning.
As a historical note, one of the first applications of distributed multisensor data fusion processing (circa 1971) was an application called HEARSAY. HEARSAY was an AI application that used a distributed agent-based blackboard architecture for speech recognition and understanding, focused on uncertainty and hypothetical interpretations.
Because of the complexity of this problem, effected by psychology, semantics, rules of discourse, syntax, lexicon, prosodic system, phonemic system, articulatory apparatus, and noise, the HEARSAY (and HEARSAY-II) problem solving model is of interest to may experts in distributed sense-and-respond applications, like CEP and the RSS/Atom use case offered by Mark Hapher at Sun.