Foldput

Foldput is a term used in programming discussions to describe a pattern that combines a fold (reduce) operation with a put-like side effect to emit results as they are produced. The aim is to process input data incrementally without building a complete intermediate collection, which can save memory and enable streaming pipelines. In a foldput, a binary combining function f, an initial accumulator init, and a sink put are supplied. As each input element x is processed, the accumulator is updated to acc' = f(acc, x) and the put function is invoked with acc'. The process continues for the entire input sequence, yielding a stream of outputs corresponding to the evolving accumulator.

Variants and semantics: Foldput can be implemented in strict or lazy evaluation contexts, with or without buffering. In strict settings, outputs are produced immediately; in buffered variants, acc' may be emitted only after a buffer fills. Handling of errors and termination may affect ordering and replay.

Applications and notes: Foldput is relevant in streaming data workflows, real-time monitoring, and memory-constrained processing. It trades off potential IO interleaving and error handling complexity for reduced peak memory usage. The concept is not tied to a particular language and is discussed as a pattern rather than a formal primitive. See also fold, map, reduce, streaming IO.