Computer Science Colloquia
Thursday, February 26, 2015
Advisor: Kevin Sullivan
Attending Faculty: David Evans (Chair), John Knight and Yanjun (Jane) Qi
Rice Hall, Rm. 504 at 9:30 AM
PhD Qualifying Exam Proposal
Potential Benefits of Synthesis from Relational Logic Specifications
The main contribution of this project is a set of experimental results that suggest that, at least in certain domains, automated synthesis of large spaces of software implementations from relational logic specifications, followed by dynamic profiling of implementation properties and trade-offs, can produce implementations that are strictly and significantly better than those produced by ordinary means. In our case, the domain is that of object-relation mapping (ORM), which is used to produce relational database schemas and interfacing software for data persistence in object-oriented software systems; and the ordinary means comprise the use of ubiquitous ORM tools. Our experiments were enabled by the design, implementation, and instantiation of a general framework for such automated dynamic synthesis and tradespace analysis, which we also contribute as a novel and useful result of this work.