Jos de Jong

It may be the holy grail of programming: writing reusable code. Not getting stuck by inventing the wheel again and again, but moving forward and building on top of what has been done. Looking at what we do today, we still haven’t really figured this out. So you are writing a software library. You had a great idea for a simple, broadly applicable solution. You have cooked up a nice API and are happy with the result. Then, new features are needed. The library grows and grows, and becomes extremely versatile. To account for more and more specific cases though, the API grows at a disproportional rate. Due to this feature creep, soon there isn’t much left from your initial, beautiful API. At some point it becomes almost too complicated to work with the library due to all configuration options and edge cases.