Invariant Rules for Software Producibility and Assurance
Final rept. 15 Mar 2009-30 Sep 2010
STATE UNIV OF NEW YORK AT STONY BROOK RESEARCH FOUNDATION
Pagination or Media Count:
This project develops a unified framework for rigorously capturing and applying software design and development knowledge to significantly improve software producibility and assurance under complex and challenging requirements facing Navy software for cyberspace. The framework is based on invariants, which underly all requirements about dependencies, concurrency, distribution, fault-tolerance, security, and general safety and correctness as well as cost and efficiency conditions. Invariant rules are used to declaratively specify how complex invariants are maintained under all possible updates to system slates. The design and development knowledge captured by invariant rules underlies not only invariant maintenance for design and optimization, but also invariant verification for validation and assurance, as well as general transformations for instrumentation, refactoring, etc. We especially investigate the use of invariant rules for specifying critical aspects of complex systems, such as in web frameworks and mashups, that may involve concurrency, distribution, trust and security.
- Operations Research
- Computer Programming and Software