Blog
Mert Özen's blog posts on software development, backend, frontend, and DevOps.
What Is the SDLC? A Software System's Life from Idea to Retirement
The SDLC (Software Development Life Cycle) is the full set of phases software goes through from idea to retirement: requirements, design, implementation, test, deployment, operations and retirement; models don't change the phases, only their order and repetition. Using Bookstore's returns feature we look at each phase's output, the weight of maintenance, the four process groups of ISO/IEC/IEEE 12207:2026 and the forgotten retirement phase.
The Waterfall Model: What Royce Actually Said in 1970
In his 1970 paper Royce drew the single-pass model but immediately called it risky, because problems such as timing and memory first appear in testing and force a redesign. We cover his five fixes (program design first, document it, do it twice, plan testing, involve the customer), the story of the name "Waterfall" and DOD-STD-2167, and how integration blows up in the last month of Bookstore's returns feature.
The V-Model: Pairing Every Development Step with a Test Level
In the V-Model testing isn't done last, it's designed on the left arm: requirements pair with acceptance tests, system design with system tests, architecture with integration tests, module design with unit tests. We tie Bookstore's 14-day return rule to its requirement with xUnit v3 traits, list the four tests of REQ-RET-01 with one command, and see in real test output which requirement a boundary bug breaks.
Incremental, Iterative and Spiral Models: Surfacing Risk Early
Incremental development adds the system part by part, iterative development builds the whole roughly and reworks it; Cockburn recommends using both together. We start Bookstore's checkout with a walking skeleton and grow it in slices, and cover Boehm's risk-driven Spiral model, throwaway vs evolutionary prototypes, RAD and RUP's four phases, together with the pitfall of mini-Waterfall disguised as iterations.
From Waterfall to Agile and DevOps: Why the Feedback Loop Shrank from Months to Hours
The feedback loop shrank as the fixed cost of deployment fell: cheaper deployments meant smaller batches, and smaller batches reduced both risk and the time to find a bug. We trace the line from the lightweight methods of the 1990s to the Agile Manifesto, DevOps in 2009 and Continuous Delivery, and cover DORA's five metrics, shift-left and the water-scrum-fall trap of running sprints while still releasing once a quarter.
Choosing the Right SDLC Model for Your Project, Team and Risk Profile
The right SDLC model is chosen by the project's risk profile, not the team's habits: requirements uncertainty, regulation, contract type, deployment cost and access to feedback. We use Cynefin to separate complicated from complex problems, build a decision table, and honestly describe conscious hybrid models for Bookstore's bank integration, subscription product and internal reporting work.