It's old. As in 16 bit compiler old. It's how you get more than just the 64K memory for your total program, IIRC. You're still limited by data segments of 64K (again, IIRC), but it's basicly so you can use extended memory beyond just the 64K.
It's been obsoleted with the appearance of 32bits (and 64, and 128.. you get the idea).
If you're using a real compiler, you don't need to use 'near' and 'far' any more. Otherwise, Google up 'near' 'far' 'pointers', and you'll get some more info.
Quzah.
