Dan McKinley's 2015 essay “Choose Boring Technology” has resurfaced as a framework for deciding when software teams should adopt unfamiliar infrastructure. Its central metaphor gives a company roughly three innovation tokens: a deliberately approximate budget for technologies whose behavior and operational burden the organization has not yet learned.
The argument does not equate old with good. McKinley distinguishes mature tools with understood capabilities and failure modes from technology that is merely dated or poor. His examples of useful boring choices include MySQL, PostgreSQL, PHP, Python, Memcached, Squid and cron. A new database, language or service-discovery layer may still be justified, especially when it is central to the company's actual product.
The cost appears when local choices accumulate. Adding a tool means monitoring it, testing it, operating it, teaching it and maintaining expertise after the original adopter moves on. A technology that is optimal for one narrow task can therefore make the entire system harder to run. McKinley's alternative to selecting the best tool for each isolated job is choosing a smaller set that occupies an acceptable position across many problems.
Long-term reliability dominates the equation because systems usually spend far more time being operated than initially built. Mature tools reduce some unknown risks because organizations already know how they fail. Newer technology contains both known questions and failure conditions teams have not yet imagined. The essay argues for making those unknowns a conscious expense rather than pretending every addition is free.
McKinley is not an absolutist. He recommends asking first how the current stack could solve a problem, writing down why that route is prohibitively difficult, and discussing additions with organization-wide visibility. When a new component overlaps an existing one, teams should commit to a migration timetable instead of leaving both implementations indefinitely.
An Etsy activity-feed example illustrates the claimed payoff. The feature was built on the company's shared PHP, MySQL, Memcached and Gearman platform rather than a more specialized addition. It later handled 20 times its earlier usage without targeted changes because improvements to the common platform benefited it. For full-text search, however, Etsy adopted Solr because forcing that task into raw PHP was not practical.
The essay's enduring message is restraint rather than stagnation. Novel technology should be spent where it advances the mission or solves a documented limitation. Familiar infrastructure, when adequate, preserves engineering attention for problems that distinguish the organization instead of consuming it on avoidable operational variety.
The framework ultimately asks a social question as much as a technical one: which new responsibilities is the whole team prepared to carry for years?
