Quote:
Originally Posted by twowheels
True, though there will most likely always remain a separation of transient vs. persistent data. It just so happens that the transient data is currently stored in the faster system RAM and the persistent data is stored on slower magnetic or flash media, but the separation is still important lest you be unable to "reboot" to clear out any state issues, to start over with a clean slate, which is often necessary due to logic bugs in software.
|
Not always a separation, I think.
When I started, we loaded programs into persistent core memory, and we had to find an "empty spot" (or find a program to "delete" from core memory) to load a new program. We kept a paper chart showing the start and length of all the programs stored in core memory, and to run one, we would set its start address on the front panel switches and press the RUN button.
If you could afford to have a hard disk on your computer, storing programs on it was the same. You needed to know the starting sector and length on disk, and an empty location big enough in core memory, then you could copy bytes between core and disk, or between disk and core, to load or save a program.
Now that flash (and other forms of nonvolatile) memory is getting so fast and cheap (as I mentioned earlier), future computers will have so much persistent fast storage that they will not need a hard drive, and RAM will be obsolete as well.
A computer with terabytes of persistent fast memory will be very nice indeed. And no need to worry about obsolete distinctions between volatile and persistent storage either...
