It has never been so hard to develop good software as today. Developers need more knowledge and skill than ever before. They need to be skilled in programming languages (e.g. Java, C++), system software platforms (.NET, J2EE), XML, middleware (WebSphere, Logicworks, etc.), the Unified Modeling Language, the Rational Unified Process, web architectures, etc. And they need to learn about these technologies faster than ever with almost no time for training and education. If they don't, their only recourse is to find shortcuts, use lightweight methodologies and ignore well-proven best practices. And as usual quality will suffer.
There is another way. In this
tutorial we will discuss how software agents can be used to reduce the gap
between the individual developers' knowledge and what is needed. For
instance, agents can minimize the process adoption thresholds so that the
complexity of a process can become transparent to the developers and thus be
perceived as lightweight. The individual developer will focus on the problem
solving and creative part, letting the agents do the work that can be guided
by formalized knowledge. We will discuss the process of formalizing
knowledge as rules, how these rules will trigger in a given context and how
the agents can propose resolutions. Examples will be used to demonstrate the
feasibility of agents in software development.
Audience:
System analysts, project leaders,
software developers, people interested in methodologies, process development
and software development tools.
Required experience:
Some experience of
software development, and of UML.
Presenter's profile: