IP Addressing is fairly easy, and if you've ever run networks on Unix, Windows, Linux, or BSD it's practically the same here.
Without going too deep into the technical inner workings of IP networks, thier size and range are determined by what are called subnet masks. Subnet masks basically tell a computer what part of the ip address holds the information about what other ip addresses are directly connected to it. In other words, for a computer to have the ip address of 192.168.1.1 and a subnet mask of 255.255.255.0 tells the computer that the ip addresses 0-255 are located on the same network segment it is, and that it doesn't need to talk to any routers to communicate with them.
To understand different subnet masks and their ranges, you need to understand that a subnet mask is a decimal representation of a binary number. In the case of the normal 255.255.255.0 subnet, it is represented in binary as: 11111111.1111111.1111111.11111111.00000000. The reason why you might want to change the subnet mask would be if you had a limited amount of IP addresses, and needed to split a range into multiple ranges which would be routable. For instance, assuming you only had 256 ip addresses (192.168.1.0-255), and you had two APs which needed to use those numbers, you'd divide create two smaller subnets. To do that, you'd need to add a 1 to the last group of 0's and convert that number from binary to decimal. In binary, the 255.255.255.0 subnet after adding a 1 would look like this: 11111111.1111111.1111111.11111111.10000000. When converted to decimal, it would look like: 255.255.255.128. The affect of doing this would allow you to create two networks, 192.168.1.0-127 and 192.168.1.128-255. You can keep doing this until you get to the subnet of 255.255.255.252. That is the smallest useful subnet, with two usable IPs and two reserved IP addresses.
A small table with the various subnets likely to be used in Star-OS networks:
|Binary||Dotted-Decimal||Slash Notation||Total IPs||Usable IPs|
You might find this calc handy - http://jodies.de/ipcalc