Domain Analysis








Next: Reverse Engineering Tools
Up: Reverse Engineering
Previous: Scope and Tasks

Domain Analysis

Current reverse engineering technology is typically based on program
analysis methods such as parsing and data-flow analysis. The domain
analysis approach has been devised as an alternative methodology
[3]. A domain is a problem area, and domain
analysis is an attempt to identify important objects, operators,
and relationships within that problem area or domain. Reverse
engineering process benefits tremendously from the knowledge
obtained through domain analysis.

Jarzabek [4] proposes a domain-model driven approach for
reengineering procedural business systems into systems having object-oriented
views. The reengineered code is piggy-backed on relational model.
Jarzabek adopts a data-driven, rather than a process-driven, approach
and this enables him to perform the task more easily, since
object-oriented paradigm is itself data-driven. Moreover, data is more
permanent than processes. The approach also takes into account two
basic properties of OO paradigm, namely pre-conditions and
post-conditions. The objective of his method is to create a
repository of reusable software components identified and
obtained through domain analysis, which is performed using pattern
matching, program slicing, and other static program analysis
techniques. Jarzabek summarizes his method as a process consisting
of ten steps, the first eight of those are essentially reverse
engineering phases.



Mohammad Ashrafuzzaman
Fri Oct 25 00:55:30 CST 1996