Lightweight program specialization via dynamic slicing

Program slicing is a well-known technique that extracts from a program those statements which are relevant to a particular criterion. While static slicing does not consider any input data, dynamic slices are computed from a particular program execution. Thus, dynamic slicers are usually easier to design and implement. In this work, we present a program specialization technique for lazy functional logic programming which is based on dynamic slicing. Our method exploits the code size reduction capabilities of slicing in order to produce a version of the original program specialized w.r.t. a given criterion. We also introduce some simple, post-processing transformations that allow us to further simplify the specialized program. The kind of specialization performed by our approach cannot be achieved with other related techniques like partial evaluation. Categories and Subject Descriptors F.3.2 [Logics and Meanings of Programs]: Semantics of Programming Languages—operational semantics Ge...

• Type: Conference
• Year: 2005
• Where: WCFLP
• Authors: Claudio Ochoa, Josep Silva, Germán Vidal
• Keywords: Artificial Intelligence, Dynamic Slicing, Functional Logic Programming, Lazy Functional Logic, WCFLP 2005