High-Level Operations in Nonprocedural Programming Languages.
MOORE SCHOOL OF ELECTRICAL ENGINEERING PHILADELPHIA PA DEPT OF COMPUTER AND INFORMATION SCIENCES
Pagination or Media Count:
Despite having many advantages, nonprocedural programming languages have limited acceptance because of the inefficiencies of their implementations, especially for high-level operations. This dissertation presents a source-to-source transformation approach for efficient implementation of high-level operations in nonprocedural programming languages. Efficiency is achieved via decomposing the high-level operations into sets of basic ones prior to their being translated into the underlying graph representation. Global optimization hence becomes possible through source language manipulation, scheduling program events, and code generation. The specific contributions of the research include a A source-to-source transformation scheme for efficient implementation of high-level operations, b Operations are made more powerful by allowing more generally structured operands, c Special indexing patterns are identified and uses for the application of efficient storage allocation schemes, d New building blocks for the underlying array graph are suggested and e The expressive power of the MODEL language is increased substantially through the incorporation of high-level operations.
- Computer Programming and Software
- Computer Hardware