Originally posted on Plio’s Blog End-to-end testing is often discussed in ideal terms — stable environments, modern browsers, and predictable user behavior. In practice, teams building software for real users operate under far more constraints. This post documents how we approached end-to-end testing while working on Plio, an open-source education…

Automating operational reporting for a large-scale education program During the COVID-19 pandemic, many education programs were forced to move online almost overnight. For NGOs running at-scale interventions, this shift wasn’t just about delivering content — it was about maintaining visibility into learning outcomes without adding unsustainable operational load. As part…

Early architectural decisions and their consequences Plio was designed to be used by multiple organizations concurrently, each running its own programs, users, and content, while relying on shared infrastructure. This was not a scaling concern discovered later. It was intrinsic to how Plio was expected to be adopted: different organizations…

Why ColoredCow treats code quality as infrastructure, not preference Most software systems are built with the hope that they will grow in importance.What differs is whether they are designed as if increasing responsibility is inevitable — and as if small decisions made early will be paid for later. At ColoredCow,…