The seminar is devoted to modern problems of software engineering, development tools and methods, program analysis and verification. Topics discussed in the seminar include
Requirements engineering, analysis and modeling
Paradigms for computers modeling
Static and dynamic program analysis
Dynamic verification and monitoring
Automatic test case generation
Verification completeness analysis
Performance modeling, measuring and testing
Security and safety analysis
Integration of verification methods
Problems of propagation of new technologies to software engineering practices
Problems of teaching software development and analysis
The seminar is intended to students, researchers, engineers, industrial experts. The companies participating in the seminar include Intel, Microsoft, Яндекс.
The seminar takes place in Moscow, Russia, each third Thursday every month. Start time is 5p.m. Auditorium “110″ of the Institute for System Programming of Russian Academy of Sciences (see map on the right).
Sergey Berezin — Ph.D., associate professor at Lomonosov Moscow State University, head of r&d group «Information Technologies in Science».
Dmitry Voytsekhovsky — graduated from Lomonosov Moscow State University, trained in Microsoft Research Cambridge, now he is RSDE at r&d group «Information Technologies in Science».
Speakers will share their experience of creating software for efficient composition and evaluation of computational experiments. They will describe their approach and set of F# components codenamed Angara. Project Angara accumulates results of their long term collaboration with Computational Sciences group in Microsoft Research Cambridge. Angara helps researchers to build reproducible computational experiments that can be performed multiple times from scratch with identical results. Full provenance information is available for every result artefact allowing to trace its origins and understand how exactly it was computed. Angara supports efficient incremental construction of computational experiment. Processing steps can be added or altering with re-computing only affected parts of the experiments. This saves significant amount of time because many computational experiments are long-running and require significant computing resources. A scientist can observe intermediate results as soon as they are produced so he can understand that something goes wrong on early iterations and doesn’t wait until computation complete.