eitherloopt

Eitherloopt is a concurrency pattern described in software engineering literature and online discussions as a construct that initiates two independent loops or long-running tasks in parallel and returns the result of the one that completes first. The term is not part of any formal language specification; it has appeared mainly in discussions of latency-sensitive algorithms and in experimental libraries.

Semantics and implementation: Each loop runs with its own cancellation scope; a coordination primitive signals completion; the winner's result is propagated; the losing loop is canceled if supported; exceptions from the losing task may be suppressed or propagated depending on design; overall semantics depend on language and library. In practice, this pattern is implemented via futures/promises and cancellation tokens; similar concepts exist as Promise.race in JavaScript, asyncio wait(..., return_when=FIRST_COMPLETED) in Python, and CompletableFuture.anyOf in Java.

Usage: When multiple strategies can compute a result with different cost profiles, eitherloopt can reduce latency by racing them. It is common in network requests, heuristic searches, or algorithmic decision-making where early results are valuable. Risks include increased resource usage, potential nondeterminism, and difficulty in error handling and cancellation leaks; the pattern adds complexity and is not suitable for all workloads.

History and adoption: The pattern traces to online communities in the 2010s and 2020s; it is not standardized. Some language ecosystems include built-in support through async runtimes, and discussions often compare it to race semantics from other languages.

See also: Promise.race, futures::select, asyncio wait, race condition.