In a Register column marking the protocol's longevity, Rupert Goodwins argues that TCP/IP's success comes down to a series of pragmatic design decisions. The most important was splitting the original protocol into TCP and IP. IP handles best-effort packet delivery between addresses, while TCP adds reliability and ordering. That separation means alternatives such as the congestion-easing Homa protocol can run alongside TCP rather than requiring a wholesale replacement.
The column credits TCP/IP's open, consensus-driven development for its victory over proprietary rivals like IBM's SNA and telecom-backed ATM. It also highlights two later fixes that kept the stack alive: Van Jacobson and Mike Karels' 1988 congestion-control work, which prevented the early internet from collapsing under its own growth, and NAT, which let IPv4 continue serving billions of devices despite its 32-bit address limit.
Goodwins acknowledges the result is not always elegant—NAT is described as "fugly"—but argues that this willingness to adopt imperfect solutions is exactly why TCP/IP still carries most of the world's traffic. The column is a celebration rather than a technical deep dive, and it does not engage with criticisms of the stack's security or performance overheads.