The general problem by xkcd
Getting started with a software product or project is messier and stubbornly more manual than I would like to think. Once started, the lure of sirens awakens: “how about generalizing this”, “refactoring that”, “come here and optimize with us!”. Then the complexity of the project rises like Gartner’s hype cycle on its first curve, and the progress grinds to a halt.
To borrow a Finnish phrase, “kävellen olisit jo perillä”: “by walking, you’d already be there.” It means carrying metaphorical buckets of water instead of designing a better way too soon. Sometimes I’m astute enough to foresee this in my own behaviour and stay my course amid chaos and, usually, that leads to finishing earlier.
The common theme here is that completionism takes exponentially more time and effort. The exponentiality results from a compound interest effect of toil and costs.
Mind the gaps and mind the complexity to prevent this exploding toil-effect. In music, I heard the advice to play below skill limits to allow room for musicality. Donald Knuth warned: Premature optimization is the root of all evil. Fournier mentioned in her book about a quote on political reform stating that good reform is one that is effective even if only half the people commit to it. That’s the kind of robustness you want in a project that disrupts and scales out: progress that survives partial adoption. In product management you don’t concentrate on a product line before having a product. In software development some say that to scale you have to do things that don’t scale.
So heed this: remember that you should have walked – most beginnings don’t need leverage, they need legs. Kävellen olisit jo perillä.