In software engineering two opposing philosophies dominate large‑scale projects. Evolutionary development starts with a minimal product and adds features over time, while big upfront design insists on a complete specification before any code is written. Proponents compare the former to a startup’s rapid iteration and the latter to constructing a skyscraper.

A former employee of a fifty‑year‑old enterprise described a landscape of more than 3,000 active systems, spanning dozens of business lines and multiple tech stacks. Inconsistent data, security gaps, and operational friction plagued the environment. Consolidating into a handful of well‑architected platforms could slash complexity by up to ninety percent, cut costs, and improve resilience.

Choosing between the two approaches hinges on how teams handle dependencies. Ignoring them early accelerates delivery but forces costly retrofits; mapping them up front slows momentum but yields a more stable architecture. Most organizations settle on a hybrid path: outline a high‑level design, iterate in sizable chunks, and refactor as hidden dependencies emerge.