How Algorithms Show That Some Jobs are Unnecessary

Reading Algorithms to Live By: The Computer Science of Human Decisions by Brian Christian, it’s easy to get immersed in the many different applications, explanations, and more that Christian provides. One such topic I’d like to talk about more is the idea of caching, and how it shows that when organizing any collection of objects, oftentimes no system at all is more efficient than a certain sorting pattern.

You’ve probably heard of caching, I believe in years past the district installed a “Clear Cache” extension that totally did nothing, but it was a cool little trash bin that said it cleared your cache! Caching is the idea that computers store pieces of information or processes that need to be frequently used. Imagine, instead of a brightness button, if you wanted to change your computer screen you had to physically go to a terminal prompt and input your commands there, like below.

That would be super unnecessary and inconvenient! Caching helps to make computers run faster, having key processes saved to improve efficiency. 

The trickiest part of this idea is, with only a limited amount of memory that your computer can hold in this “cache”, what should we keep and what should we get rid of to maximize efficiency? 

It turns out, while there are many complex solutions, like caching reactively where your storage depends entirely on what’s being used at the present moment in time, the overall most efficient strategy is a last-in-first-used strategy. Especially with a computer, information most recently accessed has the highest chance to be called again. (For more caching examples, check out: https://docs.aws.amazon.com/whitepapers/latest/database-caching-strategies-using-redis/caching-patterns.html!)

How can we visualize this? Imagine doing your math homework, let’s say your processes of completing this task are writing with a pencil, fixing mistakes with your eraser, solving problems with your calculator, and correcting mistakes with a red pen. Let’s visualize each of these objects as different processes on your computer. 

Let’s say you’re working with your pencil, solving a problem. Then making a mistake, you quickly erase it. What will you pick up next? Most likely your pencil to continue the problem. Lets say, after solving that problem, the next one requires a calculator. After picking up the calculator and solving the problem, you write down your answer with your pencil. Realizing you forgot to add a value, you quickly adjust your calculations and fix your work. 

Here, we can see that you kept picking up the school supplies that you used last – this idea of the most recently used tool being the one most likely to be used again.   

So how does this apply to our lives? 

Imagine you are a librarian, sorting books. I love this picture below with the annotations and circling marks – it perfectly gets at my claim. How should you sort books? Should the big Shakespeare book go together with the Romeo and Juliet books, even though one is a biography and the other is a work of fiction? 

This problem is what librarians face in their day-to-day operations: how should they sort these books to be efficient for readers to search for? 

Yet applying Caching concepts, and the idea of last used as the most likely accessed process, it makes the most sense for librarians to never reshelve a book. Since we know people will want to use the most recently used process, in this case, a certain book, why should librarians ever sort these books back into the shelves where it’s harder for patrons to find? Not only does this take time out of their day, but it also wouldn’t be efficient for readers. 

Which is super strange! It seems obvious that we should sort items, and order the chaos. But here, the most efficient strategy for librarians is to have a “recently checked out” book display in the lobby of the library, and only reshelve the books that have been on this display for a certain amount of time, let’s say a week. 

This realization leads us to question if reshelving books is even needed, and if not why so many spend their time doing something inefficient for both the librarian and reader. 

These ideas are super interesting to explore: it shows how some jobs in our society, or even entire processes in our society, could be unnecessary and even inefficient for society. 

Overall, I hope this blog helped show how we can use algorithms to not only better our lives, but learn a thing or two from the computers that power our society.