On Fri, Feb 16, 2018 at 4:35 PM, Larry Baker <ba...@usgs.gov> wrote: > In a true VM architecture, different objects can be distinguished by their > program addresses. This is a key distinction between memory mapping and > virtual memory >
Larry - I 100-% agree with you. My point was that from a programmer's standpoint, I can (and have as Tim describes, I used those maps too), remember pushing past the address space limits on our PDP-8 which was were I first messed with manual overlays. But we had real VM (as described by Denning) in the IBM 360/67 running TSS at the same time. Our first PDP-10's did not and of course the 11's did not either. So I just remind folks that programming in a memory constrained environment can be done by an application program to get around the limits of the number of address bits that the HW supplies. VM is supported by the OS, without application programmer intervention - which certainly makes it desirable. As Tim pointed 32 bits was both 'not enough' and the memory scheme could be abused. It's not a universal solution to all memory issues. It is handy, but we lost something too. I was suggesting that not enough programmers are taught about overlays. Frankly they should be before they learn about VM; IMO. Learning how your program lays out in memory and how it is going to perform, what it is call graph look like etc... these are skills that you acquired naturally in the old days and much hard to understand, but less realize why they are important. BTW: a similar lost art of that of mathematical interpolation and 'significant digits.' I'm the last of a breed of engineer that learned on a slide rule too; which is where you quickly get an understanding of those two concepts. That said, just as I'd rather not go back to having to do manual overlays; I did love my TI-50 (first calculator that have transcendental functions) when I got it in the early 1970s. :-) Clem ᐧ
_______________________________________________ Simh mailing list Simh@trailing-edge.com http://mailman.trailing-edge.com/mailman/listinfo/simh