Automation is meant to make our lives simpler: just throw a shell script at some problem and the problem goes away, right? If only life in a modern DevOps world was so easy! Instead, we have thousands of tools to choose from, and hundreds of problems to solve, even in small environments. Life wasn't meant to be easy, but does it have to have so many command line arguments?!?
Many of the decisions we make regarding automation are based on some combination of “experience”, “that's how it's always been”, and “I just need it to work”. This is not helpful when you have little experience, a green-field site, or even less time available than you thought. Even with experience, we tend to fall back to “Just use \<insert favorite battle-tested tool here /\>”. _We choose the tool not because it's the right tool for the job, but because it's a tool we know_. This becomes further “justified” on the grounds of personal productivity. Suddenly, tool choice is the important decision, and the original problem is secondary. How can we do better?
Mike has been working with UNIX since first gaining access to a BSD-based system in the late 1980s. After mastering the `oops(8)` and `restore(8)` commands, he gained a deep appreciation of the need to automate, and of the automation potential that is a fundamental part of UNIX and UNIX like systems.
Mike currently lives in Amsterdam with his wife, no cats, and a 3D printer named “Bob”.
