Developer Spent Two Years and 300+ Iterations Solving Reactive Programming's Core Problem

A software developer spent two years and over 300 design iterations building a new reactive programming paradigm from scratch, centered on a deceptively simple question: how do you connect a producer and consumer running at different speeds. The effort explored push-based streams, async generator-based pull systems, and their respective trade-offs around backpressure, buffering, and fan-out. Push-based approaches struggled when consumers were slow, while pull-based async generators performed poorly and made multi-consumer scenarios difficult. The developer ultimately arrived at a 'credit system' that separates consumer readiness signals from actual value delivery, decoupling the two sides of the data flow. The resulting library, published under the handle watchmanlab/stream, represents the culmination of lessons learned from repeatedly hitting architectural walls across each discarded approach.