Packet loss is taboo; to an Internet architect, it immediately signifies an inefficient design likely to exhibit instability and poor performance. In this paper, the authors argue that such an implication is not fundamental. In particular, there exist design points that provide many desirable properties including near optimal performance while suffering high loss rates. They focus specifically on congestion control, where researchers have long clung to the belief that loss avoidance is central to high throughput. A protocol that supports the sharing of resources that exist in different packet switching networks is presented.