BAD METAPHORS is an ongoing series that takes a critical look at the figures of speech that shuttle between technology and everyday life. Read the others here.


In the 1940s, when computers were the size of rooms, it took one’s entire body to program them. It demanded movement, reaching and hauling cables, adjusting complex series of switches that presented themselves as arrays in physical space. Programming, that is, was rooted in an embodied understanding: It was not merely thought but felt, through a constant confrontation with the limits of both the hardware and our own physical capabilities.

That physical effort was part of what it meant to be “close to the metal” — programmers’ jargon for computing that dealt directly with materials of machines. Instead of writing code for communication between operating systems and drivers, programmers “close to the metal” affect the hardware itself, dealing with the machine’s transistors, resistors, and other material components at the base level of the computer’s performance. Being close to the metal meant understanding the machine’s switches before they are covered by layers of abstraction. To the mostly male programmers of the era, that conveyed a sense of their exclusive dominance over machines.

As the mouse brought users into a more intimate symbiosis with the machine it also distanced them from the “metal”

In 1968, when computers were the size of desks, Douglas Engelbart invented the computer mouse as part of his “oN-Line System” (NLS). As John Markoff describes in What the Dormouse Said: How the ’60s Counterculture Shaped the Personal Computer, the audience who saw its initial demonstration at a San Francisco computer conference were surprised, in the era of mainframes, to see the system “being used interactively with all its resources appearing to be devoted to a single individual.” Mainframe computers then required separate operators for inputting data and analyzing the output. The oN-Line System reimagined computing, staging all the processes for a single user. In Engelbart’s system, the body was not reaching for the computer; the computer became an extension of the body.

As Paul Dourish notes in Where the Action Is, there “has been a long transition from interacting with computers using a soldering iron to interacting using a mouse.” As the mouse brought users into a more intimate symbiosis with the machine, offering direct command over it, it also distanced them from the metal — from the computer as a physical, material object. It brought us into a world where the computer can appear as an immersive world in itself. Users were not close to the metal but at one with it.

But this shift did not destroy the desire of some to be “close to the metal” in the sense of having exclusive or privileged access. It did, however, change its terms: Simpler interfaces meant that it no longer required soldering circuit boards to retain a sense of a superior mastery of machines. Just knowing how to code the software that sat on top of the computer’s increasingly small silicon chip meant you were among the elite. This meant that programmers retained status not by being physically closer to material but by dictating and controlling others’ interfaces.

If we can’t get into the hardware or even the software, then closeness to the metal is about getting into the consequences of using them

Now, human-computer interaction is shifting away from the mouse and toward even more intimate interfaces: Touch devices are another step toward the forgetting of the body as it merges with machine. Like the mouse but more so, touch provides the illusion of even more direct agency and therefore of being in more control. But if you are closer to the physical metal when using a laptop’s trackpad or touchscreen, you are no closer to understanding how it works or, more important, how you work on your laptop and how it works on you. “As computer interfaces increasingly shift toward touch pads and touch screens,” Ali Na warns in “The Fetish of the Click: A Small History of the Computer Mouse as Vulva,” “it is important to avoid understanding this transition as one toward more direct or unmediated contact with the digital.” The more direct the interface seems to ordinary users, the more likely they are to take it for granted and ignore how their experience is being mediated. This in turn allows the would-be programming elite to remain “close to the metal” by taking advantage of those users’ ignorance (even as coding itself becomes more automated and distant from the machine).

But closeness to the metal need not imply users becoming ever more fused with machines, and being ever more manipulated by a programming elite a few steps ahead of them. Instead, we can reconceive what “being close to the metal” could mean, return it to the sense of physicality and friction that it once implied, and update the ethos for programmers accordingly.

If we can’t get into the hardware or even the software, then closeness to the metal is about getting into the consequences. It’s about making connections between people’s lives, their bodies, and how computationally built systems have an effect on them. Closeness to the metal can signify a push back against elitism and dominance, toward a more inclusive and more inquisitive understanding of computers.

The removal of the body from our understanding of computers can seem to imply a removal of accountability from the programmer. Many digital platforms tend to urge this perspective by pushing responsibility into the hands of the user where users only have a very limited set of agencies. But remembering the mouse can be helpful here. With the touchscreen dominating our interactions, the mouse, which once seemed to grant agency (albeit through narrowly prescribing the form of our interactions), has been recontextualized. It now appears as full of friction, which gives it the potential to interrupt the perceptions of computing that it once facilitated.

Contrasting mouse use with touch screens illustrates the speed at which computing technology urges us to move, and what it has overlooked in the process. That is, mouse use can now bring us closer to the metal in the sense of encouraging us to pause. It can be a moment for the programmers to take their hands off the keyboard and consider how their code might have an effect on a real life.

This can be helpful in considering different bodies, different sorts of users — how the friction of an interface is always relative to the person using it regardless of the interface’s ultimate goal: efficiency. The mouse has potential to remind us of the body. It can, in turn, bring us closer to understanding the complex history of how computers came to be personal and very much about the bodies who brought them to where they are today.

If you are close to the metal now, then you are committed to re-evaluating complex systems in order to have a fuller understanding of them

Today, being “close to the metal” can be an acknowledgment by technologists that there is value in questioning and complicating what we think we know about the machine’s deepest workings. If you are close to the metal in this sense, then you are committed to re-evaluating and reframing complex systems to have a fuller understanding of them. It is recognizing that there is work to be done toward undoing computational systems which prioritize algorithmic prowess over people’s lives. Coding a new machine learning model, for example, takes us even one step further away from the metal, where mathematical representations of bodies have unprecedented and often negative effects on bodies. Being closer to the metal means insisting on accountability for the social implications of these systems.

This contemplation is about re-prioritizing the body over the computer, ensuring that the computer is something between bodies as opposed to in control of them. Like the reaching arms of the programmers in the 1940s, many of whom were women, the mouse can bring us back into our body as we learn to live with and through computers.

Many technologists today are “close to the metal” in this sense. Committed to unearthing these negative effects before they ever reach the human body, data scientist Rumman Chowdhury leads the group Responsible AI at Accenture Applied Intelligence. The group, which operates on the basis that technology is always directly shaped by its makers, develops tools to ensure the fairness of a data set or algorithmic process. While addressing potentially harmful technology through adjusting the technology itself is important, so is disseminating resources on existing technology’s effects. The Detroit Community Technology Project conducts research, creates educational material, and sustains initiatives which are committed to data justice and the equitable access and use of technology. They recently published “Our Data Bodies: Digital Defense Playbook,” which is meant to support groups involved “in intersectional fights for racial justice, LGBQ liberation, feminism, immigrant rights, economic justice, and other freedom struggles, to help us understand and address the impact of data-based technologies on our social justice work.” Also committed to unearthing technology’s effects on human lives, Ruha Benjamin, associate professor of African American studies at Princeton University, has presented research in a talk entitled “A New Jim Code?” which explores technology’s built-in discriminations through the “world of biased bots” and “altruistic algorithms.” Benjamin asks us to “question not only the technologies we are sold, but also the ones we manufacture ourselves” by providing information and tools which can help us to dissect the technologies that we are confronted with on a daily basis.  And there are many more examples: Joy Buolamwini, Cathy O’Neil, and Wendy Hui Kong Chun, to name only a few.

To be close to the metal, then, is considering the social implications of the code. Knowing what’s inside the tiniest of computer chips is no longer be a priority when beginning the work of deconstruction of large computational systems.