Separating parsing and analysis in reverse engineering tools

The authors describe the lessons learned in extending the capabilities of a reverse engineering tool to analyze both an additional dialect of the language it was initially built to parse and a new embedded assembly language. The effort involved in this extension provides data to support the assertion that reverse engineering tools should create a clean separation between parsing the source code and analyzing it. A language independent modeling approach is discussed that will allow achieving this separation. Additional advantages that accrue by maintaining this separation, such as multiple language support and support for design recovery, are discussed