Act as a Principal Distributed Systems Engineer. Compare the Raft and Paxos consensus algorithms in the context of building a fault-tolerant, strongly consistent distributed key-value store. Specifically, analyze their leader election processes, log replication mechanisms, and safety properties. Discuss the operational complexities and ease of implementation for each. Conclude with a recommendation for a system that prioritizes understandability and robustness over极致 throughput, and provide a pseudocode outline of the leader election logic for your chosen algorithm.