Next, there is the problem of incremental compiling. Am I correct in thinking that Perlis, with his “threaded lists,” has that problem, and the related problem of com- pile-test-recompile, essentially solved?
Over on the hardware side, I am worried that the boundry- registered problem, or more generally the memory-protection problem, may be expensive to solve on the Q-32 and both difficult and expensive to solve on other machines, and I am worried that the problem of swapping or transferring information between core and secondary memory will be difficult and expensive on 7090s and 7094s–and I worry that time-sharing will not be much good without fast swaps or transfers. What are the best thoughts on these questions? In what state are our several or collective plans?