Short note: Information in forward/reverse engineering

Information content is the amount and level of detail of design information that is available to or produced by the process.

Forward engineering

• Goes from abstract, high-level information to detailed low-level information.
• Input: requirements, use cases, domain and design models (class, sequence diagrams) with business rules and design intent.
• Output: source code and database schemas. Information is added at each step (algorithms, data types, platform details).
• Information content therefore increases as the process moves towards code, and the intent of the designer is known.

Reverse engineering

• Goes from low-level to high-level.
• Input: source code, binaries, database. Contains much detail (control flow, data structures) but not the requirements, design rationale or business rules.
• Output: recovered class diagrams, call graphs, sequence diagrams, documentation.
• Information is abstracted; some information (comments, intent, naming, design decisions) is lost or has to be guessed, so the recovered models are less complete than the original ones.

 Forward:   Requirements -> Design -> Code     (info added)
 Reverse:   Code -> Design -> Requirements     (info abstracted,
                                                 intent lost)

Round-trip engineering combines both so that the model and the code stay synchronised.