It is a matter of the memory model you are using. On old or embedded systems, some memory was outside of the range of a normal pointer. If you have 4 megs of ram you need at least a 22bit pointer to see all of it. But let's say you only have a 16 bit pointer. This means you can only access the first 65K of ram. Odd as it may sound, this was a problem on old computers, and is sometimes an issue on embedded devices with limited processing power. The near and far classifications were a solution. Pointers are near by default. In my example above, the 65K of ram would be accessed with a near pointer. To get past that 16 bit limit, you need a far pointer. Thus: memory within the pointer's range is near. Memory outside of the range is far. Near pointer: char near * ptr; Far pointer: char far * ptr;
A far pointer uses both the segment and the offset address to point to a location in memory. A near pointer in contrast uses only the offset address and the default segment. The far pointer can point to any location in memory, whereas the near pointer can only point to a nearby local address.
Something that was important 20 years ago. Now you can forget it.