Regression Verification in a User-Centered Software Development Process for Evolving Automated Production Systems (IMPROVE APS)
The vision for this project is to advance technology such that regression verification methods are available that will be routinely used for ensuring correctness in the evolution processes for long-living reliable systems requiring frequent re-validation. The goal of regression verification is to formally prove that software remains correct through its evolution, changes have the desired effect, and no new bugs are introduced.
Regression verification avoids the main bottleneck for the routine practical use of formal verification, namely the need to write full functional specifications (which is a huge effort). Also, given two program versions or variants that are both complex but similar to each other, the effort for verifying the relation between them mainly depends on the difference between the programs and not on their overall size and complexity.
As a follow-up to the IMPROVE project, we plan to overcome the two remaining obstacles on the path to reaching the above vision of routine use:
- Reach and power: The reach and power of regression verification needs to be extended to cover real-world systems and change scenarios.
- User in the loop: Regression verification needs to be integrated into the software development and evolution process, and useful feedback needs to be given to the user in case a verification attempt is not immediately successful.
In this phase, we will target a particular application domain. This provides the third main motif for our project in Phase 2 – now called IMPROVE APS:
Automated production systems in the pharmaceutical and the food-manufacturing industry:
The control software and discrete processes of automated production systems (aPS) are a very promising application area for regression verification. These systems are long-living and have to fulfill dependability and reliability criteria to avoid machine hazards or accidents involving operator or maintenance personnel. Applicability in realistic industrial situations is within reach as we focus on evolution with limited changes (e.g., bug fixes) and small variations in the software. As case studies, we use manufacturing systems in the pharmaceutical and the food industry, where aPS must be validated following regulations to ensure no harm is inflicted on consumers of the products. The envisaged solutions of IMPROVE APS follow the goal of a comprehensive elaboration of approaches to manage software evolution with additional constraints from application taken into account.
History
This project is the successor of the IMPROVE project, which focused on regression verification of imperative and object-oriented programs. We have investigated the foundations of regression verification, developed the basic methods, and shown their applicability to realistic examples and case studies.