Programming as Navigation

Programming can be thought of as a navigation problem.

I used to walk these streets. Now I travel by taxi, and whether I keep my inner map good enough depends on whether I pay attention to the routes. The taxis learned to drive in a typical city. In an unfamiliar city, each ride makes the next ride more likely to follow the previous ones. A taxi digs its pathways and makes some thicker and thicker on its map.

So now, you operate a fleet of taxis. Exciting, huh?

You need to make sure your fleet travels efficiently.

Changing or relearning the inefficient map they have developed (the codebase) is not easy or cheap. You can, of course, decide to keep logistics inefficient; that's up to you. You may decide that you are not in charge of the fleet operating efficiently at all. The logistics aspect solves itself somehow. That means you will find that some parts of the fleet are trapped in mud, have started to take some weird pills, or are cheating on you by doing something else entirely.

Is that a problem if you continue to profit from the company? If your goal is to get the maximum profit, an inefficient fleet is your problem. If you have competition, an inefficient fleet is your problem. If you want to provide good service, an inefficient fleet is your problem. We can agree that none of it matters if you are the only passenger.

We can continue to build this analogy and think of how operating the fleet of taxis can fail. Or what it means to operate that fleet well. Or what the city means here, whether it is changing, and whether there are police or rivalries. Or how much more pleasant it is to drive a taxi than to operate a fleet of taxis. We could think of a conflict between a taxi driver and a fleet manager. We could think of a conflict between multiple fleet managers.

Let's stay curious; let's continue to find ways to do good.

Comments !

links

social