1 Cloud-Based Science? Bertrand Meyer Chair of Software Engineering, ETH Zurich Eiffel Software, Santa Barbara Cloud Futures panel
2 The message The cloud changes everything Research will move from experiment-based to data-based Computational science will move from simulation to analysis Inductive approaches will get a new boost
3 Benefits for science Scale Realism Surprises Openness Reproducibility
4 Issues (a very partial list) Programming concurrent applications Reliability Security Scaling up
5 Modern science is inductive Karl Popper Gaston Bachelard Thomas Kuhn
6 Terra incognita: software engineering Empirical software engineering (mining software archives) Program verification: Model checking, SAT solving Proofs Automated testing (See our AutoTest tool, IEEE Computer, September 2009)
7 An example: CloudStudio
8 Verification as A Matter Of Course Arbiter Programmer Suggestions Boogie prover Sep. logic prover AutoFix AutoTest Test case generation EiffelStudio Suggestions Test execution Test results Interactive prover