Code Should be Readable

On the first day of my first job, at age 23, I learned the most important lesson of my life about programming: code should be readable. Four years of CS at university did not prepare me for this, but one day of work did (thanks Achiel!).

I had created something "clever" and was barked at by the Senior Engineer of the team that my solution was too complicated. For a second I thought he was suffering from low IQ, but he explained that:

  1. He was not able to instantly understand my code. Lesson: It was apparently vital that other people in the team understood my code. In fact, it was not even my code, the team owned it.
  2. The problem I was solving was simple and not worth spending much mental capacity on. Lesson: writing code is just a small part of what happens with it, code is much more frequently read, tested, debugged and refactored. All these activities take time and mental capacity.
  3. In 6 months, I would have forgotten the code and would look at it the same way he did now (that is: with a frown and raised eyebrows). Lesson: You are temporarily suffering from understanding the code too well, make sure you compensate for that.
  4. We were paid (quite well) to make product for the company, not to be smart-asses. Lesson: time is money and there is something called Opportunity Cost that makes your time even more valuable. Boring code is good code.
These lessons have stuck with me forever and made me allergic to complicated code. Of course sometimes problems really are complex, and there is big difference between "essential complexity" and "merely complicated". Senior Software Engineers should have a well-developed gut feeling to distinguish between the two.

I don't care much for fanatic discussions about Test Driven Development or Micro Services vs. Monoliths because reality is much less clear-cut. I think it's MUCH more important that whenever you create software, or basically anything*, that you keep this in mind:
  • the complexity of the solution should match the complexity of the problem
  • what you create should be easy to understand for those who work with it

As long as you build your software by these two rules, you should be alright.

* read The Design of Everyday things by Don Norman and Don't Make me Think by Steve Krug

Popular posts from this blog

Idea time: RFID+E-Ink, electronic price tags without batteries

Parsing 10TB of Metadata, 26M Domain Names and 1.4M SSL Certs for $10 on AWS

The software that I love