The way subnets work, wouldn't connecting two router interfaces together require their own subnet between them. Unless that subnet mask has 31 bits, wouldn't that was adress space. I'm asking because I often seen that done in networking books. How can this be done without wasting IP Addresses? They usually draw this when explaining subnetting. They have a central router connected to several other each one supposed to be creating their own subnet. Is this really how subnettimg is done?

Example

<-------[Router 1]-----Wasteful Subnet-------[Router 2]------>
             |
             |
             |
            \/