We should not ship code

Charles-Axel Dein

It easy to be mistaken into thinking that a software developer's job is to ship code.

It is not.

Our job is to solve intellectual problems.

It so happens that out of all the tools humans created, computers are best at solving those, because software is great at dealing with abstractions. However, we will never get recognition from the business if we just ship code. We get recognition if we solve problems and if we do it fast.

That is why the best companies focus on problem-solving in their interviews, and not on fundamental knowledge. A quick learner can learn whatever she wants to solve things. We are in the learning business. We are always learning new programming languages, databases, libraries, services, patterns... to be faster at solving problems. We also need to learn about the problems itself, and that is why waterfall is rarely a good thing: how can developers fix business problems if they cannot build their understanding of it?

Yesterday, during our show & tell session one developer demonstrated a remarkable tool that would fix one of our long-standing problems. This tool requires minimal integration. A problem was solved without code.

We are in the business of solving intellectual problems. The business is never going to measure our ROI with the number of lines of code we write. The only thing it is concerned with is whether or not the problem was solved efficiently.

It does not mean that everything has to be simple, nor this is a good excuse for fast and loose solutions (ultimately those will make the code less maintainable, and it will be very costly to solve new problems, as requirements tend to grow over time). We should take pride in elegantly solving complex problems. Some complex problems might require a complex solution. That is fine. Just like woodworkers prides themselves in their mastery of their tools, we should be proud of how much we master our tools. However, our skills are not the end goal. The solution is.