Were “Windows-like” GUIs explored in the ENIAC era (late 1940s–early 1950s)?

 ENIAC (completed in 1945) and other machines of the late 1940s to early 1950s were operated mainly via punch cards and physical switches, with output on lamps or printers. They did not have window systems or mouse-driven GUIs like modern Windows. However, several projects from that period did explore screen displays and pointing devices—early ideas that foreshadowed later GUIs. Below is a chronological overview of key papers, reports, and experiments (English and Japanese literature).

1945: Vannevar Bush’s “Memex” (essay As We May Think)

• Citation / Author / Year: As We May Think — Vannevar Bush, July 1945 (The Atlantic).

• Summary: Bush proposed a future personal information machine called Memex: a desk-like device holding vast literature on microfilm with multiple on-desk viewing screens. A user would operate keys, buttons, and levers to retrieve items and display related materials side-by-side, building “associative trails” (an early hypertext concept). While not implemented, it anticipated screen-based interaction and user-driven information manipulation, influencing hypertext and personal computing.

1946: Ralph Benjamin’s trackball—the first pointing device

• Citation / Author / Year: Patent filing for a “pointing device” (UK), 1947 — Ralph Benjamin; invention in 1946.

• Summary: For Royal Navy radar work, Benjamin devised the “Roller Ball”—the first trackball—to smoothly control a cursor on a radar screen by spinning a metal ball. Although classified and not commercialized then, it presaged the ball mechanism later used in mice. In 1952, the Canadian Navy’s DATAR project built a large trackball (even using a five-pin bowling ball) and successfully field-tested it in 1953. Thus, mouse-like pointing originated in the late 1940s/early 1950s.

1948: Manchester “Baby” (SSEM) and CRT display output

• Citation / Author / Year: SSEM project reports, University of Manchester, June 1948.

• Summary: The SSEM—first stored-program electronic computer—displayed memory contents as bit patterns on a CRT. This is among the earliest examples of a computer using an electronic screen for output. The display showed dots for 0/1; it wasn’t directly interactive, but it planted the seed for later graphical output.

1948–1951: MIT “Whirlwind I,” interactive CRTs, and the light pen

• Citation / Author / Year: Project Whirlwind reports — Jay Forrester et al., ca. 1945–1951.

• Summary: MIT’s Whirlwind I emphasized real-time operation and used CRT displays. Engineers (e.g., Robert R. Everett) developed the light pen, letting users point at luminous spots on the screen and have the computer detect that location—arguably the first true pointing input to a computer display. By 1949, experiments included simple screen-based demos (moving light points) at around 256×256 resolution. Douglas T. Ross later recalled experiments where a user “wrote” directly on a horizontally mounted oscilloscope—an early gesture toward today’s pen/touch interaction. This marked a move from standing at switch panels to sitting and interacting with on-screen objects.

1954 onward: SAGE—console displays and light guns at scale

• Citation / Author / Year: SAGE (Semi-Automatic Ground Environment) reports — MIT Lincoln Lab, IBM et al.; development started 1954 (operational by 1963).

• Summary: The USAF’s SAGE defense network, derived from Whirlwind, deployed large AN/FSQ-7 computers and numerous operator consoles with large CRTs and light guns (light pens). Radar targets appeared as symbols on screen, and operators selected them with the light gun to issue commands—a practical, at-scale graphical interface. Those target symbols functioned much like today’s icons, and selecting them with a light pen anticipated point-and-click.

Early–mid 1950s: Other proposals and experiments

• Douglas T. Ross (mid-1950s): At MIT, Ross advocated direct access (interactive keyboard input and immediate response instead of batch card submission). A 1956 memo and subsequent Whirlwind experiments connected a Flexowriter to enable typed, conversational use. Ross also explored light-pen and handwriting concepts—proto “personal workstation” ideas of sitting with a screen and collaborating with the machine.

• “Stylator” (1957): Tom Dimond and colleagues demonstrated real-time handwriting input, anticipating later pen computing (e.g., RAND Tablet, 1964).

---

Bottom line: The late 1940s to early 1950s did not produce a WIMP GUI (windows, icons, mouse, pointer) like modern Windows. But they did produce crucial precursors:

• Screen displays (CRT) used by computers,

• Pointing devices (trackball, light pen) and on-screen selection,

• Multi-document viewing concepts (Memex’s side-by-side screens and associative trails),

• The shift toward interactive, real-time, sit-down use.

These paved the way for the 1960s breakthroughs—Ivan Sutherland’s Sketchpad (1963) and Douglas Engelbart’s mouse and windowed demo (1968)—that crystallized the GUI.

Selected references (by era):

• Vannevar Bush, “As We May Think,” The Atlantic, July 1945 — the Memex concept (hypertext precursor).

• Ralph Benjamin, UK patent filing (1947) — earliest trackball (“Roller Ball”).

• University of Manchester, SSEM (Baby) reports, 1948 — CRT memory display.

• Jay Forrester et al., Project Whirlwind reports, ca. 1945–1951 — real-time computing, CRTs, light pen.

• Robert R. Everett et al., MIT Lincoln Lab, ~1950 — light-pen development (carried into SAGE).

• SAGE program (MIT Lincoln Lab, IBM), 1954–1963 — large-scale light-pen graphical consoles.

• Douglas T. Ross, recollections and memos (1950s; later reprinted in ACM History of Personal Workstations, 1989) — early “personal workstation” viewpoint.

• Tom Dimond et al., 1957 — “Stylator” handwriting input.