Abstract

Introduction Excessive requirements for memory space have in the past hindered or even prevented otherwise attractive applications of functional programming. Although this could be blamed in part on space-hungry implementation methods, in most cases it would have been possible to cut memory requirements very significantly by making a few changes to the source program. But there were no tools to help programmers make appropriate changes. Usage of heap memory was reported only as a total volume of allocations; there was no way to investigate how different parts of a program made demands on heap memory --- something which may be far from apparent in the source of a sizable program making use of lazy evaluation and higher-order functions. Finding the appropriate place to make a space-saving change could be very difficult. In the last few years, there has been a renewed effort to provide appropriate profiling tools for functional programmers. Because the first aspect of performance

Citations