LSR ASE 2005 Panel on Education in Automated

LSR ASE 2005 Panel on Education in Automated Software Engineering Yves Ledru LSR/IMAG, University of Grenoble-1, (France) Long Beach, CA, Nov. 11 th 2005

The french educational system LSR Master 1 st year Master 2 Pro Master 2 Research Ph D Software Engineers (practitioners of ASE) Researchers (designers of ASE) Should we teach the same ASE courses to both kinds of students?

Automatic generation of applications from a UML diagram LSR Delphia Object Modeler GROUP PERSON http: //www. si. fr. atosorigin. com/dom/ Groupcode Groupofperson Members Lastname Groupname • 1 Level: Master 1 * Firstname • Context: UML+Z course Cardnb Tel [*] • Objectives: – Show that the concepts learned in the course are at the basis of such tools – Experimentation with such tools – Limits of such tools : show that the specification of constraints on the database schema is shorter than their actual implementation with deamons

What should software engineers learn? LSR Master 1 st year Concepts Practice the tools Master 2 in order to see: Pro The magic of ASE tools My experience in this curriculum: Master 1: DOM The limits of ASE tools Master 2: Db. C executable Software Engineers specifications (practitioners of ASE) (VDM, JML)

Research : KIDS/VDM LSR • Theme: prototyping VDM specifications using KIDS • Underlying concepts: arn o le t! s t ec ing roj th ny r 2 p – Design by contract (VDM) ma ste oo ma – Program synthesis (KIDS)T or a • Underlying technologies: f – Theorem Proving (KIDS) – The Refine language: Functional, Declarative, imperative, Objectoriented + Input/output grammars!

What should Ph. D students learn? LSR Master 1 st year Concepts Bad news: Underlying techniques are very dependent on the kind of tool you are trying to build! Underlying Master 2 Techniques! Research Ph D Evaluation techniques! Researchers (designers of ASE)

What should they learn? (Merging) LSR Master 1 st year Concepts Practice the tools Master 2 in order to see: Pro The magic of ASE tools The limits of ASE tools Software Engineers (practitioners of ASE) Underlying Master 2 Techniques! Research Underlying Ph D Evaluation Techniques! techniques! Researchers (designers of ASE)

Use your own medecine! LSR • Tool: TOBIAS • Combinatorial testing of Java programs • Tool developed by 2 Ph. D students + 2 Master students • Tool written in Java • Succesfully used outside of our team • But… today never used to test TOBIAS itself!

What should they learn? (Merging) LSR Master 1 st year Concepts Practice the tools Master 2 in order to see: Pro The magic of ASE tools The limits of ASE tools Software Engineers (practitioners of ASE) Underlying Master 2 Techniques! Research Underlying Ph D Evaluation Techniques! techniques! Researchers (designers of ASE) Researchers should be the first actors to bridge the gap between research and practice!

LSR