The PDP-8 was a weird segmented-address-space machine; you had to solve a "knapsack problem" to fit subroutines into the memory. http://en.wikipedia.org/wiki/Knapsack_problem On modern architectures, your programs will run w/o solving a knapsack problem, but if you want the best performance using a cache, you might still want to solve a knapsack problem. At 05:21 PM 2/8/2011, David Wilson wrote:
My first programming experience was in FORTRAN on a PDP-11 in 1977. As I remember, the PDP-11 did not have automatic paging, so if your program was too big for memory, you had to write overlay files to tell it where to load its subroutines. If two subroutines were in different overlays in the same memory space, one could not call the other, and recursive subroutines were verbotten.
When I was a freshman at Cornell in 1978, I had to submit my PL/I homework programs on punch cards to the IBM mainframe (properly prefixed by pretty pink JCL cards), where they were allotted a generous 15-second runtime, with a limited amount of output spewing out on oversize computer paper on the industrial size impact printer. But the fortunate few got to work on VAXen with amber text monitors running VMS or UNIX. I still have a Digital VAX coffee mug.