Loading field map
Human-Computer Interaction (HCI) began with a deceptively simple question: how can we make computers usable by people? But from the start, researchers disagreed about what "usable" meant and who should decide. Some treated the user as a cognitive system with measurable limits; others argued that good design should feel like manipulating physical objects. Still others insisted that users themselves must shape the technology. Over four decades, these competing visions produced a family of frameworks that continue to coexist, each offering a different lens on interaction.
The first systematic framework for HCI was Human Information Processing (HIP), articulated by Card, Moran, and Newell in their 1983 book The Psychology of Human-Computer Interaction. HIP modeled the user as an information-processing system with limited attention, memory, and processing speed. It produced quantitative predictions—for example, how long a user would take to find a command in a menu—and aimed to make interface design an engineering discipline grounded in cognitive psychology. For about a decade, HIP dominated academic HCI research, especially at institutions like Xerox PARC and Carnegie Mellon.
Almost simultaneously, a very different approach emerged. Direct Manipulation, championed by Ben Shneiderman in the early 1980s, argued that interfaces should let users act directly on objects rather than through abstract commands. A direct-manipulation interface makes actions visible, reversible, and incremental—think of dragging a file icon to a trash can instead of typing "delete filename." Where HIP treated the user as a cognitive bottleneck to be optimized, Direct Manipulation treated the user as a skilled actor who could learn through exploration. The two frameworks competed directly from 1985 to 1995, with Direct Manipulation gradually winning over the broader design community because it offered intuitive principles that matched how people actually learn and use software. Yet HIP did not disappear; its legacy persists in computational cognitive modeling (used in human factors and usability testing) and in the idea that interface efficiency can be measured and predicted.
While American HCI focused on individual cognition and interface style, a different tradition was taking shape in Scandinavia. Participatory Design (PD), emerging from the late 1980s, grew out of workplace democracy movements and union-led projects. PD argued that users are not just sources of requirements but co-designers of the technology they will use. Unlike Direct Manipulation, which gave users control over the interface, PD gave users control over the design process itself—through workshops, prototyping sessions, and ongoing collaboration. This framework challenged the expert-driven ethos of both HIP and Direct Manipulation, insisting that design decisions are political and that those affected by a system should have a say in its creation. PD coexists with other frameworks today as a methodological stance: it does not prescribe a particular interface style, but it shapes how design is conducted, especially in community settings, public-sector projects, and research through design.
By the early 1990s, HCI had largely focused on the single-user desktop computer. A new wave of frameworks pushed beyond that boundary. Context-Aware Computing, introduced by Schilit, Adams, and Want in 1994, asked how applications could adapt to the user's location, identity, nearby people, and activity. Instead of assuming a fixed workstation, context-aware systems used sensors to infer the situation and adjust behavior—for example, a phone that silences itself in a meeting. This framework shifted HCI's attention from the interface to the environment, treating context as a resource that could be sensed and exploited.
Ubiquitous Computing (Ubicomp), famously envisioned by Mark Weiser in 1991, derived from the same impulse but aimed much higher. Where Context-Aware Computing focused on adaptive applications, Ubicomp imagined computation woven into the fabric of everyday life—hundreds of invisible computers embedded in walls, furniture, and clothing, so seamlessly integrated that the computer itself disappears. Ubicomp absorbed Context-Aware Computing's sensing infrastructure and broadened it into a vision of calm technology. Both frameworks remain active: Context-Aware Computing powers modern mobile and IoT applications, while Ubicomp continues to inspire research in smart environments, wearable computing, and the Internet of Things.
By the late 1990s, practitioners and researchers began to feel that usability—the central concern of HIP and Direct Manipulation—was too narrow. User Experience Design (UXD) emerged around 1998 as a framework that broadened HCI's goals to include emotion, meaning, aesthetics, and brand. UXD subsumed Direct Manipulation's principles of clear feedback and direct control, but added layers of pleasure, delight, and narrative. A product's user experience is not just whether it works efficiently, but how it feels to use, what it says about the user, and how it fits into a larger life. UXD also absorbed insights from Context-Aware Computing: a good experience depends on the user's situation, not just the screen. Today, UXD is the dominant framework in industry, shaping how companies design websites, apps, and services. It is less a single theory than a broad orientation that draws on psychology, design, and marketing, but its core claim—that usability alone is insufficient—has become a near-consensus in HCI.
Around 2001, a more radical challenge arrived. Embodied Interaction, articulated by Paul Dourish in Where the Action Is, drew on phenomenology and ethnomethodology to argue that meaning in interaction arises not from mental models but from physical, situated action. Embodied Interaction rejected the cognitive focus of HIP and even the screen-centric directness of Direct Manipulation. Instead, it emphasized tangible computing, social practice, and the way people make sense of technology through their bodies and their environment. This framework shares with Participatory Design a commitment to situated, social practice—both see interaction as embedded in real-world contexts—but Embodied Interaction goes further by grounding that insight in a philosophical account of how meaning is created. It has influenced research in tangible user interfaces, whole-body interaction, and collaborative systems. Embodied Interaction does not replace UXD or Context-Aware Computing; rather, it offers a different starting point: instead of asking what the user experiences, it asks how experience is constituted through action.
No single framework has won HCI. Instead, the field holds several in productive tension. User Experience Design leads in commercial practice, providing a language for designers to talk about emotion and brand. Embodied Interaction drives academic research on tangible, social, and physical computing. Context-Aware Computing and Ubiquitous Computing underpin the mobile and IoT industries. Participatory Design remains a vital methodology for projects that prioritize democratic values. Direct Manipulation has been largely absorbed into UXD as a set of interface patterns (drag-and-drop, direct editing), but its core insight—that users learn best by doing—still informs interaction design. Human Information Processing lives on in human factors engineering, cognitive modeling, and usability metrics.
What do these frameworks agree on? Nearly all accept that design should be user-centered, iterative, and evaluated with real people. They share a commitment to making technology serve human needs rather than the other way around. Where they disagree is on what counts as the user's experience, how much of interaction can be modeled computationally, and who should participate in design decisions. UXD tends to treat experience as a subjective state to be measured; Embodied Interaction treats it as an enacted process to be understood ethnographically. Context-Aware Computing assumes that context can be sensed and modeled; Participatory Design warns that context is negotiated and political. These disagreements are not signs of weakness but of a healthy, pluralistic field that adapts its frameworks to different problems, settings, and values.