Automatic Coding: Choice of Data Structures

1. Front Matter

   Pages N2-vii

   PDF

2. Introduction

   • James Richard Low

   Pages 1-15

3. Overview of the System

   • James Richard Low

   Pages 16-21

4. Abstract Data Structures and their Representations

   • James Richard Low

   Pages 22-35

5. Information Gathering

   • James Richard Low

   Pages 36-44

6. Selection of Data Structures

   • James Richard Low

   Pages 45-50

7. Results

   • James Richard Low

   Pages 51-57

8. Conclusions and Directions for Future Research

   • James Richard Low

   Pages 58-62

9. Appendices

   • James Richard Low

   Pages 63-98

10. References

    • James Richard Low

    Pages 99-103

11. Back Matter

    Pages 104-110

    PDF

1.0 2 information structures, and the runtime efficiency available using lower level data structures. We believe these runtime inefficiencies can be remedied by providing a library of representations for the information structures and a user assisted automatic system to choose appropriate elements from this library for the user's programs. The main problems in such an automatic approach are: 1). Building a library of representations; 2). Characterizing the properties of these representations; 3). understanding how the user's program actually uses the abstract information structures; and -4). efficiently matching the user program needs with the properties of the known representations. The research reported here is an attempt to demonstrate the feasibility of such automatic representation selection. We allow the user to express data in terms of relations, sets, and sequences. Our system then obtains information by analysis of the program, execution statement profiles and interrogation of the user. Using that information, the system selects efficient (in terms of time and space) low level implementations for the user's information structures from a fixed library of information structure representations. Though we have considered only a few abstract information structures (those available in our programming language, SAIL) we believe the techniques used in this system are generally applicable.

• character
• coding
• data structure
• data structures
• efficiency
• implementation
• information
• language
• Matching
• programming
• programming language
• SAIL
• sets
• system
• time

Policies and ethics