ICSE 2011 Panel - Lionel Briand

9
© Lionel Briand What Industry Wants from Research ICSE 2011 Lionel C. Briand Simula Research Laboratory and University of Oslo

description

 

Transcript of ICSE 2011 Panel - Lionel Briand

Page 1: ICSE 2011 Panel - Lionel Briand

© Lionel Briand

What Industry Wants from Research

ICSE 2011

Lionel C. Briand Simula Research Laboratory

and University of Oslo

Page 2: ICSE 2011 Panel - Lionel Briand

© Lionel Briand

Perception from industry

•  Disconnect

•  Scalability (heuristics versus exact methods)

•  Applicability (context factors and constraints)

Page 3: ICSE 2011 Panel - Lionel Briand

© Lionel Briand

Compelling evidence

•  Realistic conditions –  Scale –  Human factors –  Impact of change –  Cost-benefit analysis

Page 4: ICSE 2011 Panel - Lionel Briand

© Lionel Briand

Research questions

•  Many … and they have not changed in 20 years

•  Requirements changes

•  Large scale testing

•  Product lines and complex configuration

•  Certification

Page 5: ICSE 2011 Panel - Lionel Briand

© Lionel Briand

Most relevant empirical finding

•  Requirements traceability

•  Regression testing

•  Global software development

•  Change propagation (MSR)

•  … ???

Page 6: ICSE 2011 Panel - Lionel Briand

© Lionel Briand

Success Stories

•  (Model-based) testing and empirical research at Microsoft Research

•  Research on inspections and model-based testing at Fraunhofer

•  Our research at Simula, though recent, is having a significant impact on our partners

Page 7: ICSE 2011 Panel - Lionel Briand

© Lionel Briand

Dissemination to industry

•  Different research paradigm than the dominating one is needed Problem

identification

Problem

formulation

Study

State-of-the-art

Candidate

Solutions Initial

validation

Realistic

validation Release

solution

Ind. Partners

Lab

Page 8: ICSE 2011 Panel - Lionel Briand

© Lionel Briand

Engage industry in research •  Engage researchers in industry: Get

researchers to take an interest in industry problems

•  Different research paradigm (engineering research)

•  Combine research and technology transfer (same people)

•  Understand problems an assess solutions in context

Page 9: ICSE 2011 Panel - Lionel Briand

© Lionel Briand

Reasons and Solutions for the Perceived Lack of Impact of SE Research

•  Potential reasons: –  Scalability –  Interest –  Publications –  Difficulty of real problems –  Integration of research results

•  Partial solutions –  Different research paradigm –  Scalable solutions: combining MDE & SBSE –  Empirical studies: better understanding, demonstrable result –  Long term relationships: understand the domain, focus on

important problems