Data over Code

An interesting video that made think about how I approach problems. This was an interesting perspective and an excellent communicated talk.

The key takeaway is recognize that this approach we are looking to identify and label side effects. We identify which are the side effects.

Need to identify the parts of your code:

  • Data required for your program/system
  • Computations over the data
  • Actions are the side effects

When you compose the componenets you “pollute” your system.

Steps in workflow are:

  1. Get it to work
  2. Identify the side effects, and push it to the edge of the system
  3. Reify your system. Convert your function into a data structure.

Take code, make it into data. Make functions to manipulate such data strcutures.

In this python video this princple is done and exactly what should be done for projects.