In many ways, computers mirror the inner workings of our brain. As computer scientists tackle resource challenges in these machines, there are parallels we can draw about our own minds. One of these interesting problems is how to be as responsive as possible.
In their book Algorithms to Live By, Brian Christian and Tom Griffiths explain that computers store data in a few different places (p. 85). Fast memory is the fastest to access, as the name implies. It’s also the most expensive to produce so there’s relatively little built into each computer. Standard memory is next quickest and second most-expensive. Finally, computers can write data to the disk. This is the slowest of the three areas to retrieve data from, but is also least expensive so there tends to be lots of it.
When it comes to making a computer responsive, the faster you can get data to the processor for use, the sooner the user gets the result they want. Ideally, you could keep everything in fast memory, but as mentioned above, it’s too expensive. The best case is if you could guess what the user is likely to need next and keep it in fast memory and move less-likely data into standard memory (p. 88). Anything referenced only occasionally would be written to disk. The problem is how to anticipate what the user is going to need next.
It turns out that studies have shown that what a user is likely to need next, is something he or she has needed in the recent past. When people are working on problems, we tend to use the same bits of information several times before we no longer need it. So, computers use this discovery to sort where to keep data – most recently used going in fast memory and older data going to standard memory or disk.
In our own brains, fast memory is like short-term memory. It’s quick but limited. Psychologist George A. Miller did extensive research into how many things human short-term memory could hold at one time (Chip Heath and Karla Starr: Making Numbers Count (Avid Reader Press, 2022) 17-18). His answer was seven, plus or minus two. Further, a single “thing” can occupy multiple memory slots if the item itself is complex. For example, difficult fractions (193/436ths) and multiple-digit numbers (682,103 or 0.4023415) can consume the lion’s share of the memory slots on their own. Once short-term memory gets full, anything else is hard to keep track of or gets lost almost immediately.
Your best way to address this limit is to write things down. Once you put something on paper (or computer file, etc.), you no longer have to maintain it in short-term memory and your overall “cognitive load”, or mental work required, goes down. Once you get done with what you’ve written, you can save it in a filing cabinet for future reference, which is like writing the information to the hard drive. Using these techniques for reducing your cognitive load is especially helpful in brainstorming.