Our view of De SIRE Politecnico di Milano

brief presentation of the group • background in – software engineering – development of

a few remarks on the state of the art • why is industry often

Our view of system modeling and V&V • There is a urgent need for

Our view of system modeling and V&V • The above scenario can be nicely

Slides: 5

Our view of De. SIRE Politecnico di Milano contact persons: Angelo Morzenti: morzenti@elet. polimi. it Dino Mandrioli: mandrioli@elet. polimi. it

brief presentation of the group • background in – software engineering – development of critical systems – formal methods (FMs), V&V – temporal logic (we originated the TRIO logic) • emphasis on industrial application of FMs

a few remarks on the state of the art • why is industry often “reluctant” to adopt FMs? – they are difficult to use (notation, tools) • one needs a Ph. D in Th. Comp. Sci. to use them profitably – description and tools (V&V algorithms) do not scale – FMs are too “revolutionary” not “evolutionary” • very (too) pervasive on the development process • upset previous habits and procedures • Approach to “correct by construction” systems often rejected when based only on descriptive notations (logic, axioms etc. ) – because of existing strong cultural gaps • A viable alternative: combine descriptive and operational (i. e. , state-transition based) notations

Our view of system modeling and V&V • There is a urgent need for and integrated method and tool environment supporting seamless, methodical development from requirements to maintenance • Such environment should include – Architectural notations to model environment, components, parallelism, distribution – state/transition-based notation(s) for modeling components, to support • validation (e. g. , simulation) • design • implementation (possibly automatic) – assertion-based notation(s) for describing properties of components, to support • verification (model checking, deductive analysis) • test case generation from models • . .

Our view of system modeling and V&V • The above scenario can be nicely embedded into UML 2 – OMG standard to be released in the near future • provides a precise semantics that is necessary for definition of methods and tools • Many useful results from the FMs field can be usefully incorporated into UML 2 -based method and tool environment for the development of dependable critical systems