I talk about cognitive load a lot (over 2 dozen times in the last year). Especially about reducing cognitive load. But what is it, and why is it important? And what does it have to do with software development, especially architecture and design?
To give an analogy, let’s imagine the brain is a tool. A general purpose tool that, with some instructions/adjustment can do lots of different things, You can store those instructions away and get them back when you need to do what they’re for. This tool can only make contact with the environment around it in a few ways, but those ways can be combined for more flexibility. One tool that meets those requirements is a computer. You’ve probably heard of them.
So, the brain is like a computer. That’s a nice analogy to use to help understand cognitive load. Especially the CPU part. Consider the CPU. It’s got a few registers. It can act on those registers in a handful of ways, including getting data into and out of a register. And that’s it. Everything else it does is a combination of those things. Let’s say your CPU has 5 registers. You can do anything you want with the info in them, but if you need to work with more than 5 things you’ll need to keep stopping to put one of those pieces of info down somewhere safe, pick up the new one, and move on. The bigger the difference between the number of pieces of info and the number of registers the more time spent just moving things around, not doing anything with the info. And every time you need to move something in or out of a register there’s a chance to get interrupted, or worse, drop something and lose it.
In a related fashion, computers appear to do multiple things at once. But in reality, for a given CPU that’s not really true. It does one thing for a few milliseconds, switches to a new thing for a few more, and cycles through everything over and over again, giving the appearance of doing all of those things at once. We call the time spent between doing different things a context switch, and they can take orders of magnitude longer than actually doing the work because the computer needs to put all of the info in those registers somewhere safe, then bring back the info that was in them the last time it worked on the other thing. It also needs to remember exactly where in the list of steps it was, and pick up where it left off. Again, that context switch is great opportunity to get something wrong.
Now, your brain isn’t a CPU, but let’s stick with the analogy. There are only a limited number of things you can keep in active memory at once. If the number of things you need to remember is higher than that you have to keep refreshing your memory. That’s cognitive load. The less time you spend refreshing your memory, the more time you can spend on what you’re trying to do.
Similarly, When you’re working on something, in the zone, as it were, you’ve got all the things you need in working memory and most of the instruction cycles in your brain are going towards getting that something done. When it’s time to context switch you need to save all that state, find and reload the old state. Until that’s done you aren’t making progress on anything. And for our brains that process is very slow and imprecise. Often you can’t get back to where you were, you can only get to a recent save point and then you need to go over the same things again to get back to where you were. That’s more cognitive load. Again, keeping it down helps you make progress against your goals.
So that’s what I mean by cognitive load and why it’s important. How it relates to development is a whole set of different topics for the future.