Design Environments
ESPRIT Framework IV LTR 26241
Summary
|
The PROSPER project will research and develop the
technology needed to deliver the benefits of mechanised formal specification
and verification to system designers in industry. Examples of the next
generation of CAD and CASE tools will be produced, incorporating
user-friendly access to formal techniques. An open proof architecture
created to underpin these tools will provide the basis for other innovative
design tools in the future.
Objectives and Approach.Modern system design must satisfy increasing demands for dependable products of high quality, with shorter design times and early error detection. Incremental improvements to conventional design methods are not enough. More powerful techniques based on tool-supported formal methods are essential. They enable analysis at higher levels of abstraction, with the result that designers can check a wider range of properties than with conventional methods.Existing proof tools need experts in logic and are not integrated into established design flows. PROSPER aims to develop the technology needed to overcome this barrier. We have the concrete objective of producing two prototype design tools, one for hardware and one for software, each providing user-friendly access to formal techniques. These will be demonstration models of the next generation of CAD/CASE tools, which will support timely design of high-quality, reliable systems. The two design tools will be built by integrating theorem proving technology into existing systems. One will link theorem proving support to the VDM-SL (CASE) toolbox of our industrial partner IFAD; the other will provide proof support for the industry-standard (CAD) languages Verilog and VHDL via a common hardware verification workbench. PROSPER will develop a common theorem-proving infrastructure for these two platforms, along with the know-how to engineer its integration into other CAD/CASE tools. A further objective of PROSPER is to deliver practical knowledge of how best to feed the benefits of these next-generation tools into conventional design practice. This will be achieved by evaluating the prototype design tools on end-user applications, driven by typical customer needs faced by our industrial partners. Results.The specific results of PROSPER will be:1. prototype design tools: an enhanced VDM-SL tool for software, and a VHDL/Verilog tool for hardware; 2. user-friendly interfaces: for requirements, using natural language and timing diagrams, for proof via a GUI; 3. an extensible open proof architecture: a core proof engine, easily integrated with other tools via an API; 4. technology transfer: application case studies, publications, workshops and an exploitation plan. Scientific articles will also be published describing these results and including design principles for building the next generation of CAD/CASE tools incorporating rigorous theorem-proving technology. The novel scientific contribution of these results will be the mechanisms to introduce formal reasoning into CAD/CASE systems, based around a new open proof architecture, incorporating an industrial-strength proof engine supported by efficient plug-in proof tools and new specification interfaces. Impact and Exploitation.Producers of software and hardware in all application areas need more advanced tool kits to design current and future products. Advanced techniques are not only desirable for competitiveness; they are increasingly mandatory for safety critical applications.The strong industrial need among end-users has resulted in the identification of this requirement by commercial tool builders, including our industrial partners IFAD (Denmark) and Prover Technology AB (Sweden). These two companies exemplify the producers of tools who need the results of this project. First, producers of CAD/CASE tools (like IFAD) have identified a need for incorporating reasoning support into design tools to increase quality. Second, there are companies (like Prover Technology) who successfully sell theorem proving technology focused on a restricted class of problems. In the next generation of their tools, more general proof support is needed. In addition to laying the foundation for new European CAD/CASE tools and enhancements to existing ones, PROSPER's results will open up new possibilities to add value to CAD/CASE tools originating outside the EU. These add-ons will provide new capabilities, resulting in competitive advantage to European industry. Partners.PROSPER includes participation by the Universities of Glasgow (theorem-proving) Cambridge (Theorem-proving, Verilog tools), Edinburgh (natural language interfaces), and Tübingen (previously Karlsruhe) (VHDL tools). These academic sites will be joined by two companies-IFAD in Denmark (VDM CASE tools) and Prover Technology in Sweden (theorem proving), who provide contact with end-users and will exploit the results.
|