B Master SE National Centre Mathematics & Computer Science - - PowerPoint PPT Presentation

Software Analysis And Transformation

Organising a research team around the research software around the research team in software engineering Motivation, experiences, lessons

Jurgen J. Vinju

B

SWAT - SoftWare Analysis And Transformation

• National Centre Mathematics & Computer Science
• Programming languages and systems
• Algol, Python
• ASF+SDF, Rascal
• MonetDB, W3C
• Software Improvement Group (spin-off)
• Software Quality Assessment & Monitoring
• Reverse Engineering
• CWI SWAT ≡ INRIA ATEAMS
• all about source code
• supporting the tasks of programmers to {make,maintain} better software, cheaper
• understanding and controlling software complexity
• Master Software Engineering @{Universiteit van Amsterdam, VU, HvA}

SWAT

Master SE ATEAMS

SWAT - SoftWare Analysis And Transformation

Research team(s)

HR

Research field

Societal Impact

Education Funding

Research methods Team collaboration

External collaboration

SWAT - SoftWare Analysis And Transformation

Software Analysis & Transformation

code = data pragmatic undecidable infeasible creative, innovative multi-faceted domain- dependent

SWAT - SoftWare Analysis And Transformation

Software is wonderful, literally amazing. Our imagination is our only real constraint.

6

Understand this little piece of code

[Kees van der Laan]

7

Analysis

8

Analyse

9

Understanding software

• Simple code
• Simple question
• Complex domain-specific analysis
• Answer: absurd art

[Penrose,Escher]

the source code of ls

ls.c 3894 lines 77 pages 541 choices

“its complicated”

SWAT - SoftWare Analysis And Transformation

Research team(s)

People

Research field

Societal Impact

Education Funding

Research methods Internal collaboration

External collaboration

12

Research Tools Industry Rascal

“valorisation” (application) experiment (empirical methods) knowledge exchange (necessary domain knowledge) enable more and better tools

PL SE € €

13

surprisingly, it works!

results in different domains

PL SE Forensics Finance High-tech

education support project + contract funding community industry uptake fun ATEAMS/SWAT very well evaluated

14

15

A research team is not a software team

16

fewer resources = more investment in efficiency (accidentally) sold as research output

17

team members prioritise for academic output which is not software

18

“win-win” situation needs weekly recalibration short-term goals has highest priority while long-term tactics enables it all

19

seniors take long term perspective: maintenance, engineering documentation, UX, teaching material community management

20

giving more control to phd students:

• failure modes
• learning from students
• team spirit

21

taking losses:

• throwing away code
• weekends and holidays
• rotten low hanging fruit
• [illusion of] self-interest

22

We want (even) more: intuition for software by society patience with solving hard problems interaction with domain experts empirical evidence

• ut-of-the-box thinking
• nline social interaction

23

We need less: code (surprise!) competition for attention commercial rationalisation

24

Summary: we take more risk we burn more energy but we have more impact we make more mistakes and have more problems but also more solutions and more fun; we create.

25

Research Tools Industry Rascal

€

http://www.usethesource.io http://www.rascal-mpl.org