The Myth of the 'Waterfall' SDLC
software-developmentwaterfallsdlcagilehistorymethodology
Abstraction: Historical debunking of the rigid frozen-waterfall SDLC mythology
Key points:
- No primary source — not Royce (1970), Bennington (1956), nor DOD-2167 (1985) — actually advocates the rigid "frozen" sequential waterfall commonly attributed to them; all three explicitly describe iterative processes
- Royce's actual model includes five enhancements over a basic phased process: preliminary program design first, heavy documentation, build a prototype ("do it twice"), rigorous testing, and formal ongoing customer involvement
- The "frozen waterfall" label is traced to a 1976 Bell & Thayer paper, but the fully rigid interpretation appears to have originated from sales pitches for Agile tools needing a straw-man, not from practitioners
- DOD-2167 explicitly states "software development is usually an iterative process" and allows contractors to tailor the process and deviate from top-down order
- The historical context (expensive compute time at $500/hr vs. $1.50/hr engineering labor in 1950s) explains the emphasis on upfront analysis and documentation
- Author's conclusion: there is no universally applicable methodology; planned vs. agile approaches should be chosen based on project context
Connections: Waterfall Methodology · Software Development Lifecycle · Agile Development