Distributed consensus evolved from Paxos's mathematical rigor to Raft's design-for-understandability philosophy. Paxos, while theoretically elegant with its two-phase prepare/promise and accept/accepted protocol, has a notorious reputation for being difficult to understand and implement correctly. Raft explicitly prioritizes understandability by decomposing consensus into leader election, log replication, and safety properties, with a strong leader model that simplifies reasoning. Both algorithms tolerate (n-1)/2 failures and achieve equivalent theoretical guarantees, but Raft's prescriptive approach has led to more correct implementations and faster developer onboarding. Modern variants like Flexible Paxos and Multi-Raft address specific performance needs, while leaderless approaches like EPaxos remove bottlenecks. The key insight: an understandable algorithm that teams can implement correctly is more valuable than a theoretically pure but opaque one. The best consensus algorithm isn't determined by mathematical properties alone, but by the intersection of correctness, understandability, and practical implementation concerns.