The impact of automated memory management on software architecture

Much of the current programming effort for conventional computers is spent in planning the allocation of primary and secondary memory during the execution of a program. Virtual memory systems were introduced in order to free the programmer from much of the time-consuming and routine work associated with memory management. However, when a program requires more memory locations than are available in a particular virtual memory, the virtual memory must be managed by using virtual access methods. For these applications, the virtual memory advantage of increased programmer productivity is sharply reduced.