Everybody talks about "business rules", and how to extract them from code.  The definitions vary wildly, and the procedure to extract them are largely informal.  This confuses everybody about the nature of business rules and what exactly happens as they are extracted.

This talk will be a synthesis of ideas from the program analysis, reverse engineering, model extraction and business rules extraction community.  We will discuss the concept of abstraction as a unifying principle that ties these ideas together in a coherent framework, showing how decision tables, BPEL-style notations, models and domain-specific languages are all variations on a theme.

We will discuss the kind of technology that is required to enable the analysis of code by tracing information flows from system inputs through code to system outputs, and reverse engineering from code idioms with interactive guidance of the process by a business rule analyst.  As a case study, we will discuss how we were able to extract reliable models of a factory control process from extremely low level code for Dow Chemical industrial plant controllers using pattern matching technology to recognize common code idioms and design choices.

Finally, we will discuss how this technology is likely to evolve for use in the broader business rule community.

What Attendees Will Learn:

• What are abstractions, and how are they related to business rules.
• Show how computer code implements abstractions
• Show how information flow and pattern matching can be used to support extraction of business rules and abstraction