Hewre is some more detail. In the olden days 1960 - 1985, there were several types of physical memory: core - wired in and the fastest and probably the smallest total size in bytes; magnetic drum - a rotating magnetic drum containing maybe 1-10 kilobytes of space; magnetic tape - largest capacity but slowest access time. When I was supporting virtual memory development, we classified physical memory by speed from core (fastest) to tape (slowest). Then algorithyms were developed to move blockes of memory say 512 words into core fdrom the slower memory. These blocks were clled pages. The algorhithm wpuld decide bassed on looking down the instruction list fro a particlar process, what page contained most of the instructions and the data on which they needed to operate. This increased the overall process speed by insuring that a given instruction would have to out to say mag tape to fetch back the one piece of data it needed.
One algoritthm I rmember is Least Recently Used (LRU) where if, in looking down the instruction list, a given instruction now in core would not be called again for many cycles, it would be put back in tape.
Nowadays, we have three memory types: core, cache, and disk (tape is still used for archival store). The virtual memory idea still holds and one can have in the large core memory pieces (pges) of many diffewrent applications. As an example, the other day I had the following apps avaiable: Word, Excel, PowerPoint, EndNote, Canvas, and Maple 10. These were about 20 GB in size but I had a 3.5 GB core memory. So the paging algorithm moved data and instrauctions in and out of core with no noticble speed slowdown.
I hope this is helpful as my wife says when someone asks me a simple technical question, I respons with a complete lecture in quantum mechaics ( I did this on our first date and she almost gave up on me).
MamaMoose
