• EECS3342 System Specification and Refinement Lecture Times:
  • Class TR 11.30am, SC211.
  • Every Tuesday 13:00 LAS 1002A. Labs attendance is required and there will be required labs and Labtests during some of the sessions.

• Important: Subscribe to: 3342 Forum for the latest announcements. Please ask all questions relating to the course material on the forum (not via email). For all other questions, see me during office hours.

• See bottom ↓ of this page for login with your Prism password. Slides are available from the SVN repository (see link in the sidebar, once you have logged on).

• The Labs to be done each week are available at Labs. Lab0 must be completed in the first week of classes.

• Read the course outline regularly. It is important to read all the required readings (not all of which are discussed in class). It is important to do the suggested exercises.

• In the SVN → Docs, you will find the Rodin 2.8 Manual. Please familiarize yourselves with all the material in the manual, during the first week of classes.

• Suggested Text: Jean-Raymond Abrial, Modeling in Event-B: System and Software Engineering, Cambridge 2010. Available in Steacie. See SVN for notes.

Theory and tools for specifying computer systems (sequential, concurrent and embedded). Specification (via set theory and predicate logic), modelling, abstraction, refinement and formal reasoning are undertaken before code development so that systems are correct by construction under the stated assumptions.

Prerequisites: General prerequisites (CSE2011 3.0 and a cumulative GPA of 4.5 or better over all completed major computer science courses and CSE1019 3.0), and MATH 1090

This course provides students with an understanding of how to use mathematics (set theory and predicate logic) to specify and design correct computer systems whether the systems are sequential, concurrent or embedded. The course stresses both the underlying theory as well as the ability to use industrial strength tools that can be applied in practice. User requirements are formalized via an abstract mathematical model that is amenable to formal reasoning long before any programming activity is undertaken (e.g. as done in Event-B, Z and VDM). Successive models are like blueprints in traditional engineering disciplines and their mathematical nature allows us to reason about and predict their safety properties.

After successful completion of the course, students are expected to be able to:

• Understand the nature of formal methods and evaluate their suitability.
• Understand user requirements documents and the distinction between environmental constraints as opposed to functional and safety descriptions.
• Construct high level, abstract mathematical models of a system (consisting of both the system and its environment) amenable to formal reasoning.
• Use set theory and predicate logic to express functional and safety properties from the requirements as events, guards, system variants and invariants of a state-event model.
• Understand how to use models to reason about and predict their safety and liveness properties.
• Construct a sequence of refinements from abstract high-level specifications to implemented code and the proof obligations for showing that a concrete system refines the abstract system.
• Know the theory underlying state-event systems, refinements and their proof obligations.
• Apply the method to a variety of systems such as sequential, concurrent and embedded systems.
• Use practical tools for constructing and reasoning about the models.
• Compare the theory with classical Hoare Logic and the Dijkstra weakest precondition calculus .

Chapters 1, 2, 3, 4, 5, 9 and 15 from the text Modeling in Event-B: System and Software Engineering, Jean-Raymond Abrial (Cambridge) makes this a possible textbook for the course.

This course is 3 hours of instruction per week as well as 1 hour of supervised labs per week.

The lab time is used to give students detailed exercises and instruction in using a practical verification tool (such as Rodin for Event-B) to accompany the material in the lectures. Tools are essential to using the theory and methods on larger examples and require expert knowledge of the use of automated theorem proving methods. Students will use such tools to prove the examples that are discussed in class as well as larger examples.