Sketchpad

Sketchpad
Information
Type	Interactive drawing program
Industry	Computer graphics, computer-aided design
Developer	Ivan Sutherland
Supported Devices	TX-2 computer at MIT Lincoln Laboratory
Release Date	January 1963 (PhD dissertation)

Sketchpad is an interactive drawing program written by Ivan Sutherland for his Doctor of Philosophy degree at the Massachusetts Institute of Technology (MIT), described in the dissertation "Sketchpad: A man-machine graphical communication system", submitted in January 1963. It ran on the experimental TX-2 computer at MIT Lincoln Laboratory. A user drew directly on the computer's display with a light pen and used push buttons to issue commands such as "draw", "move" and "erase", so that apart from text labels no written language was needed to operate it.

Sketchpad is one of the first graphical user interfaces and one of the first interactive computer-aided design (CAD) systems. Sutherland received the 1988 ACM A.M. Turing Award "for his pioneering and visionary contributions to computer graphics, starting with Sketchpad, and continuing after." Sutherland later expanded from Sketchpad's two-dimensional graphics to three-dimensional graphics and a head-mounted display at Harvard University, a system often cited as an early milestone in virtual reality.

Reviewed 27 September 2026. Every thesis-derived fact checked against the Cambridge TR-574 full text; CHM, IEEE Spectrum, IEEE Computer Society, Design World, Road to VR, Crossref DOIs and the 1965 and 1968 Sutherland papers re-opened for their cited claims. About review dates.

History

Sutherland held a B.S. from the Carnegie Institute of Technology (1959) and an M.S. from the California Institute of Technology (1960) when he began the work at MIT. His first contact with the TX-2 came in May 1961. During the summer of 1961 he wrote a curve tracing program and developed early ideas about interlaced and "twinkled" displays. Late that summer a dormant project to connect an ink-on-paper plotter to the TX-2 was revived; Sutherland kept a side interest in the plotter hardware, and the plotter later inked every drawing in the thesis.

Claude E. Shannon agreed to supervise the drawing work as a thesis project in the fall of 1961. Steven A. Coons of the Mechanical Engineering Department and Marvin Minsky of the Artificial Intelligence group served on the thesis committee, and Douglas T. Ross of the Electronic Systems Laboratory advised on the "n-component elements" used to represent drawings. Wesley A. Clark and Jack L. Mitchell made the TX-2 available, and by the time of writing Lincoln Laboratory had provided about 600 hours of TX-2 time.

Sutherland's first light pen drawing program was working in early November 1961. It used "strong conditions", so that a line could be drawn parallel or perpendicular to another line but kept no record of that relationship; Sutherland wrote that this "in effect provided the T-square and triangle capabilities of conventional drafting." He also noted that Welden Clark of Bolt, Beranek and Newman had shown him a similar program on the PDP-1 somewhat earlier. In early December 1961 Shannon visited the TX-2 and suggested adding circles. When Shannon asked for paper to sketch a drawing he intended to enter into the computer, Sutherland realized that the strong conditions, which simulated conventional drafting tools, were not adequate for computer drawing. The "ring structure" data representation was in use by 1 February 1962, and a second-generation program added explicit constraints and automatic constraint satisfaction.

Sutherland described the work in the dissertation submitted in January 1963, with Shannon listed as thesis supervisor. He presented a paper of the same title at the AFIPS Spring Joint Computer Conference in Detroit in May 1963, where his affiliation was given as consultant to Lincoln Laboratory.

Hardware

Sketchpad depended on the TX-2, an experimental computer built at Lincoln Laboratory in 1956 that used high-speed switching transistors and large magnetic core memories. Sutherland credited the machine's 70,000-word memory, 64 index registers, flexible input-output control and its supply of toggle switches, shaft encoder knobs and push buttons for how quickly ideas could be tried; a push button register was installed at his request. Because the program ran only on this customized machine, Sketchpad had limited distribution as software, and its influence came mainly through the dissertation, a film of the program in use and the conference paper.

Component	Detail as described in the thesis
Computer	TX-2 at MIT Lincoln Laboratory; 70,000-word memory, 64 index registers
Display ("scope")	Ten bit per axis electrostatic deflection system, displaying spots at a maximum rate of about 100,000 per second at any one of slightly more than a million positions
Display file	Spot coordinates stored one per 36-bit word: 20 bits for position and 16 bits for the address of the drawing element that produced the spot
Light pen	Photodiode and transistor preamplifier in the pen housing, connected by coaxial cable; sensitive only to the initial blue flash of each spot, not the yellow afterglow
Pen tracking	Tracking cross redisplayed about 100 times per second; maximum tracked pen speed about 20 inches per second
Other controls	Push buttons for commands, toggle switches for functions, shaft encoder knobs for rotation, scale and position
Output	Pen plotter that inked drawings on paper

The light pen reported to the computer whenever a displayed spot fell within its field of view. Sketchpad followed the pen by drawing a cross-shaped pattern and checking which of its spots the pen saw. Moving the pen faster than the tracking cross could follow made the program lose track of it, and Sutherland used that loss of tracking as the "termination signal" that ended a drawing operation (a quick flick of the pen).

Features

In the introductory example of the thesis, the user points the light pen at the screen and presses "draw" to create a line segment that stretches "like a rubber band" from its starting point to the pen. Returning the pen near an existing end point makes the new line "lock on" to it exactly. Circles are drawn by placing a center point and then sweeping an arc. The Design World overview of CAD history lists rubber-banding, dynamic move, rotate and scale, and transparent zoom and pan among Sketchpad's editing features.

Sutherland identified three general capabilities as the core of the system: a subpicture capability for placing arbitrary symbols in a drawing, a constraint capability for relating drawing parts "in any computable way", and a definition copying capability for building complex relationships from simple ones.

Master drawings and instances

Any collection of lines, arcs and previously defined symbols could be defined as a symbol and reused as many times as the user wished. A change to a symbol's definition appeared at once wherever that symbol was used. The Computer History Museum describes the same behavior: "Changing a master object template changed all instances of its use." Instances could contain other instances, and the thesis records that an early bug made nested instances rotate in the wrong direction when the outer instance was rotated.

Constraints

Sketchpad stored the topology of a drawing explicitly. Moving one vertex of a polygon moved both adjacent sides, and lines attached to a symbol stayed attached when the symbol moved. The user applied conditions with the light pen and push buttons; for example, pointing at two lines and pressing a button made them parallel. Constraints were drawn on the screen so they could be erased or changed, and combinations of constraints could be saved as a composite constraint and applied in one step. The program satisfied constraints either by relaxation (repeatedly adjusting variables toward a least-mean-squares fit) or, where possible, by a "one pass" method that found an order in which the variables could be solved directly.

Data structure and display

Drawings were stored in a "ring structure" that let Sketchpad process topological information "with no searching at all." Sutherland proposed the terms "hen" and "chicken" for the two ends of each ring link. Deleting an element recursively deleted everything that depended on it: deleting a point removed the lines that ended on it.

The drawing was held in internal "page" coordinates separate from display coordinates, and the screen acted as a window onto the page. With a magnification range of 2000, a user could work on a 7-inch square portion of a drawing about 1/4 mile on a side. To reduce flicker, spots could be shown in an interlaced order, or in random order; Sutherland reported that random ordering produced a "twinkling" picture that avoided flicker entirely but lost small detail.

Demonstrated applications

The thesis reports that Sketchpad was used for electrical, mechanical, scientific, mathematical and animated drawings. Examples in the dissertation include repetitive hexagonal patterns, mechanical linkages that could be moved to study their motion, dimension lines, circuit diagrams, and truss and arch bridges for which Sketchpad computed the distribution of forces in the members. A drawing of a girl named "Nefertite" could be made to wink by swapping which of three types of eyes sat on her face, and a second face, traced from a photograph, changed character when its mouth was made larger. Sutherland suggested that producing the "almost identical but slightly different frames" of a cartoon semi-automatically could justify the system economically.

Sutherland concluded that computer drawing was worthwhile only when the user got "something more out of the drawing than just a drawing", such as the behavior of a linkage or the loads in a bridge.

Sketchpad III and three dimensions

The final chapter of the thesis states that Sketchpad's methods "generalize nicely to three dimensional drawing" and that work had begun on a "Sketchpad Three" to let users communicate solid objects to the computer. That system, by Timothy Johnson of MIT's Mechanical Engineering Department, was to let users draw directly in three dimensions and aim at points in a 3D drawing through two-dimensional perspective views on the display. Johnson described it in an MIT report, "Sketchpad III, three dimensional graphical communication with a digital computer", dated May 1963, and in a paper at the same 1963 Spring Joint Computer Conference as Sutherland's.

The thesis also mentions a program by Lawrence G. Roberts that recognized simple objects in photographs well enough to produce three-dimensional line drawings, which Roberts stored in Sketchpad's ring structure so that the results would be compatible with the three-dimensional version.

Influence

In the preface to the 2003 electronic edition, Alan Blackwell and Kerry Rodden of the University of Cambridge Computer Laboratory describe Sketchpad as "one of the most influential computer programs ever written by an individual." They credit it with anticipating direct manipulation conventions such as clicking a button to select a visible object and dragging to modify it, and they describe the light pen as a predecessor of the mouse. They trace its influence through David Canfield Smith's Pygmalion to the Xerox Star workstation, whose designers wrote that "Sketchpad influenced Star's user interface as a whole as well as its graphics applications."

The same preface notes that Sketchpad's class and instance-based inheritance predated the Simula programming language by several years, and that Alan Kay traced the origin of Smalltalk partly to receiving a copy of the Sketchpad thesis at the same time as a Simula distribution tape. The IEEE Computer Society's pioneer biography of Sutherland lists a display file for screen refresh, a recursively traversed hierarchical structure for modeling graphical objects, recursive methods for geometric transformations, and an object-oriented programming style among the techniques Sketchpad introduced. The Computer History Museum calls it "generally recognized as the first computer-aided drafting (CAD) program."

Connection to virtual reality

Sketchpad worked only in two dimensions, but Sutherland returned to it when writing about the future of computer displays. In 1965, while at the Information Processing Techniques Office of ARPA, he published "The Ultimate Display", which argued that a display connected to a computer is "a looking glass into a mathematical wonderland." In that essay he gave Sketchpad's "constraints" as an example of concepts with no previous visual form that a computer display can show, and ended with the idea of a room in which the computer controls the existence of matter. In a 2016 Road to VR interview article, computer scientist Fred Brooks is reported as saying that Sutherland had set out in that 1965 speech a vision of computer graphics that would look, sound, move, interact and feel real.

At Harvard, Sutherland's group moved from Sketchpad's two-dimensional graphics to three-dimensional line-drawn scenes and a head-mounted display that showed different views of a scene as the user moved their head; IEEE Spectrum and the Computer History Museum note that this project is frequently cited as an early milestone in virtual reality. Sutherland's 1968 paper "A head-mounted three dimensional display" describes spectacles containing two miniature cathode ray tubes that presented transparent "wire frame" line drawings, with a perspective image that changed as the user moved. The Computer History Museum dates the display, built with help from Sutherland's student Bob Sproull, to about 1967 and notes that it was nicknamed the Sword of Damocles because of the large beam needed to support its weight.

Publication history

Date	Publication	Notes
January 1963	PhD dissertation, MIT	Supervised by Claude E. Shannon
May 1963	AFIPS Spring Joint Computer Conference paper, Detroit	Conference paper of the same title; Sutherland listed as consultant, Lincoln Laboratory
September 2003	University of Cambridge Computer Laboratory Technical Report UCAM-CL-TR-574	Electronic edition with a new preface by Alan Blackwell and Kerry Rodden; proof-read by Sutherland

The Cambridge edition marks the original page breaks and does not correct errors in the original typescript, to serve as the textual equivalent of a facsimile.

See also

• Ivan Sutherland
• Sword of Damocles
• History of virtual reality
• Spatial computing
• Computer graphics