Graphical User Interface
A graphical user interface, or GUI, is a way of operating a computer by looking at and manipulating pictures on a screen — windows, icons, menus, buttons and a pointer — instead of typing commands. Rather than remembering that a file is copied with a particular command and syntax, a person drags a visible icon to another visible place, and the system draws the result immediately, so the possible actions and their effects are visible and learnable by exploration.
The graphical interface was invented at Xerox's Palo Alto Research Center in the 1970s, building on Douglas Engelbart's earlier work. Engelbart's 1968 demonstration at the Stanford Research Institute showed the mouse, windows, hypertext and collaborative editing. At PARC, the Xerox Alto, built in 1973, was the first computer to combine a bitmapped display, a mouse and a windowed desktop, and the Smalltalk environment developed there contributed overlapping windows and object-oriented programming. The commercial Xerox Star of 1981 added icons and a desktop metaphor with documents, folders and a printer shown as pictures, but its price limited its sales, and the ideas reached the mass market through others.
Apple and Microsoft carried the concept to ordinary buyers. Steve Jobs and Apple engineers visited PARC in 1979, and the resulting Apple Lisa of 1983 was the first mass-market computer with a GUI, though its price kept it niche. The Apple Macintosh, introduced in January 1984 with a famous Super Bowl commercial, made the style affordable, backed by detailed human-interface guidelines that insisted applications behave consistently. Microsoft released Windows 1.0 in 1985, the Atari ST and Commodore Amiga offered rival GUIs, and Windows 95 in 1995 made the graphical desktop the default for most of the world's computers. In the Unix world, the X Window System, created at MIT in 1984, separated the drawing server from application clients, which is why graphical programs can be displayed remotely.
The design vocabulary of the classic GUI is summed up as WIMP: windows, icons, menus and a pointer. Windows let several tasks share one screen, icons stand for files and programs, menus present commands as a visible list rather than something to recall, and the pointer, driven by a mouse, gives direct manipulation of on-screen objects. These ideas rest on the desktop metaphor, in which the screen is a tabletop with documents and folders, and on widgets such as buttons, checkboxes, sliders and text fields, which toolkits like Win32, Cocoa, Qt, GTK and the web's HTML elements supply ready-made. Because widgets behave the same across applications, a person who has learned one program can approach another with confidence, which is the real source of the GUI's accessibility.
GUI programs are written in a different style from batch programs: they are event-driven. Instead of running from start to finish, an application sets up its windows and then waits in an event loop, responding to clicks, key presses, timers and window changes by calling handler functions that update internal state and trigger a redraw. A handler that takes too long freezes the window, which is why responsive applications push slow work, such as network calls or model inference, to background threads. Modern interfaces on phones and tablets move beyond WIMP, replacing the pointer with touch and gestures, and rendering is often done on the GPU, but the event-and-redraw structure remains.
The GUI's strengths come with limits that explain why it did not replace the command line. It excels at discoverability and at tasks that are inherently visual, such as editing an image, laying out a page or inspecting a plot, but it is slow to automate: a sequence of clicks cannot easily be repeated a thousand times, versioned or run unattended on a remote server, which is why servers, cloud machines and many data-science pipelines run headless, with no graphical interface at all. The common belief that a GUI is simply an easier version of a command line is also incomplete, since it is a different trade-off between recognition and recall, not a strict improvement.
Read a graphical user interface as an application of the idea that seeing beats remembering: windows, icons, menus and a pointer turn commands into visible objects, an event loop turns each click into a response, and the price of that visibility is that anything not drawn on the screen is harder to reach and to automate.