Shape-based programming: sharpening the tools
Funding: 2000: $56,000
2001: $58,000
2002: $60,000
Funding or Partner Organisation: Australian Research Council (ARC Large Grant)
Start year: 2000
Summary: Our existing specialist languages, FML and FISh1, use shape-based tools to improve programming expressivity and execution speed, respectively. These traditionally antagonistic benefits will be combined in the general-purpose language FISh2, setting a new standard in language design. Also, this combination is the key to the adoption of parallel computing by non-specialists. Maintaining efficiency across many architectures requires automatic program optimisation, which must be guided by a model of execution costs. GoldFISh will extend FISh2 with a shape-based cost model to set a new standard in portable, efficient, data parallel programming, and be useable by general scientists and engineers.
Publications:
Jay, CB 2004, 'The pattern calculus', ACM Transactions on Programming Languages and Systems, vol. 26, no. 6, pp. 911-937.
View/Download from: Publisher's site
Jay, CB, Lu, HY & Nguyen, QT 2004, 'The Polymorphic Imperative: a Generic Approach to In-place Update', Electronic Notes in Theoretical Computer Science, vol. 91, pp. 195-211.
View/Download from: Publisher's site
Jay, CB 2001, 'Distinguishing data structures and functions: The constructor calculus and functorial types', Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 2044 LNCS, pp. 217-239.
FOR Codes: Application packages, Programming Languages