Enhanced Sharing Analysis Techniques: A Comprehensive Evaluation

Title: Enhanced Sharing Analysis Techniques: A Comprehensive Evaluation

Research Question: Can the precision of sharing analysis techniques be improved by combining them with other methods and tools?

Methodology: The researchers evaluated several techniques for enhancing sharing analysis, including integrating it with freeness and linearity information, using Pos (a domain of positive Boolean formulas), incorporating explicit structural information, implementing a reduced product of Sharing and Pos, reordering bindings in the computation of the abstract mgu, proposing a new mode to record ground or free variables, using linearity to improve analysis, and recovering hidden information in the combination of Sharing with freeness information.

Results: The experimental evaluation showed that combining Sharing with Pos, explicit structural information, and the reduced product of Sharing and Pos improved the precision of the analysis. Additionally, reordering bindings, the new mode for recording ground or free variables, and using linearity to improve the analysis also provided useful insights.

Implications: The results suggest that these techniques can significantly enhance the precision of sharing analysis, leading to improved performance in areas such as AND-parallelism, occurs-check reduction, and the computation of other program properties. Furthermore, the findings indicate that these techniques can be applied to other areas, such as Inductive Logic Programming, to speed up the refinement of inductively defined predicates.

Link to Article: https://arxiv.org/abs/0401022v1 Authors: arXiv ID: 0401022v1