The necessity of software abstraction

By Matt Layman on January 24, 2017

Tell me if you’ve heard this one before: unless you understand how a software abstraction works, you will not be able to use it well. I’ve heard variations of this declaration on a number of podcasts and other tech blogs and media. I think this idea is misguided. To the contrary, embracing software abstractions requires you to accept ultimately that you will not understand them all.

Our lives are full of abstractions. Without them, we’d have an extremely hard time functioning, because we’d always need to explain things from first principles. Take something as primary as blue. What if you always had to describe it as “the electromagnetic radiation oscillating at 650 x 1012 oscillations per second?” Even that decomposition of blue into more rudimentary terms is full of life’s abstractions.

Why then does this sentiment about deep understanding of software abstractions as a requirement gain such traction?

Consider a web developer working in a web framework. How far down do we expect that person to have an intimate understanding of what is going on? To display an HTML table in a browser, do we expect them to know how the browser will process a URL, build an HTTP request, transmit the request through numerous layers of network protocols, be received by a server, that likely handles the data in a language that uses a byte code interpreter, that makes operating system calls, that get translated into hardware interrupts, that are handled by numerous layers of hardware, that ultimately are logical gates, built out of transitors, whose material properties permit flow of electrons through various parts of a circuit? Last time I checked, when I added a CSS gradient to a page, I did not need to be aware of the electron flow through the transitors inside of my laptop. Where do we draw the line? What is “good enough?”

To be clear, I’m not saying that you should skip learning how software (or hardware) abstractions work. In the quest for a deeper understanding of your profession, you should absolutely try your best to understand the software you interact with. The reason I get so bothered by this idea that abstractions must be understood before they can be used well is that it is a disservice to new developers. Perhaps to a seasoned software veteran, stating that abstractions need to be understood is a trivial thing, because that veteran already has the years of applied learning. To a new developer, this burden of understanding might feel like a bludgeoning over their head. How do you not know this yet? You’re not qualified to do this until you get it. What do you mean you don’t know all the details of this protocol? I suspect it leads to feelings of inadequacy in their own experience and likely pushes many away from the field altogether.

I think all this means that we must become comfortable in our own intellectual shortcomings. It also means placing a certain degree of trust in this collectively awesome mind construct that we call software. Embracing software abstractions let’s us thrive at boundaries. It permits us to function without loading the entire context of the universe in our heads.

If you want to chat about this with me, I'm @mblayman on Twitter.



Matt Layman

Matt is the lead software engineer at Storybird.

Always eager to talk about Python and other technology topics, Matt organizes Python Frederick in Frederick, Maryland (NW of Washington D.C.) and seeks to grow software skills for people in his community.