Sketchpad III, a development of Sutherland's ideas by Timothy Johnson from MIT Lincoln Laboratory allowed the operator to draw and visualize objects in three dimensions. The display CRT was divided into four views: top, front, side, and perspective. The work reported in "Sketchpad III" by Timothy Johnson will let the user communicate solid objects to the computer. Johnson is completely by passing the problem of converting several two dimensional drawings into a three dimensional shape. Drawing will be directly in three dimensions from the start. No two dimensional representation will ever be stored.

		Sketchpad III is only capable of manipulating straight line, "wire frame," figures in three dimensional space. Neither program writing nor a knowledge of computers is required to operate the system. The definition, construction, and manipulation of three-dimensional surfaces are not included at present; hence edges which are normally hidden by forward surfaces are not obscured as they should be. Since all edges are visible, one views a "wire frame" with no covering. Explicit information about the topology of the part is stored as it is sketched. Parts of an object (lines or endpoints) can be moved in space without erasing. All attached lines will follow the moving part. Output: Visual Presentation - Graphical images of three dimensional objects are displayed on-line on a standard cathode ray tube. Because the screen is two-dimensional and the objects are three-dimensional wire frames, several viewing conventions were adopted to aid in visualizing the object in three-dimensional space. Strophic displays and similar methods of creating space sensations were not considered because of clumsiness and because of bilateral communication problems.
TX-2 Console Area - Directly below the CR1 display of the letters "3-D (ceiling-light reflections show) are four digital shaft encoders used to modify the display.

Four views of the object are displayed by program, one in each quadrant of the CRT screen (All figures in this paper describing 3-D objects were drawn with Sketchpad III and plotted off line, by machine.) A perspective view of the object appears in the upper right quadrant, and three orthogonal views in the remaining quadrants: top view--upper left, front view--lower left, and side view-- lower right. Wireframe objects displayed in a single-two-dimensional view without perspective fail to convey depth information. Perspective gives the illusion of three dimensions by supplying the familiar convergence of lines as they recede from the viewer.

A single perspective view of an unfamiliar object does not convey visually all the correct information either; for example, are the receding lines parallel or do they actually converge? Hence at least one other complementary view is necessary. Three protectively related orthogonal views were chosen for the complementary function. There are several reasons for this choice: a) the three views completely describe a straight line object in three dimensions. b) Three ninety-degree rotations of the part are simultaneously in view, reinforcing depth perception. c) Many users of Sketchpad III are uncomfortable sketching in perspective and would prefer the orthogonal system used by most draftsmen.
		Wrought Iron Chair Design.

A light-pen is used to guide drawing on the face of the CRT. The light- pen is a photo-diode mounted in a pen-like housing which is connected to the computer. A lens system in the pen housing focuses light on the photo diode giving a field of view of approximately one-half inch when the pen is held with in three inches of the CRT screen. The pen, which acts only as a receiver, responds if a scope phosphor is intensified within its field of view and interrupts the computer momentarily. Suitable programming determines which point causes the response. Thus, an existing line in a drawing can be singled out for program examination by merely pointing at the line with the light-pen.

		The light-pen also controls the position of a light spot on the scope. This in turn controls the position of a three-dimensional point permitting the drawing of new lines. The program anticipates the two-dimensional position of the pen on the scope and displays a series of dots in the form of a cross at the guessed position. By noting which points in the cross fall outside the pen's field of view, the program calculates the position of a new cross that is closer to the center of the moving pen's viewing field. Thus the two-dimensional position of the pen is tracked by a light source. The light-pen permits the CRT screen to pass information in two directions; the CRT is simultaneously both an input and an output device.
Segment of a Radar Dome.

Pushbuttons are provided which enable the operator to direct the computer program; for example, to erase what the light-pen is pointing at, to move what the light-pen is pointing at according to subsequent pen motions, to start drawing a straight line where the pen is pointing according to sub sequent pen motions, to translate the drawing according to pen motions, and so forth. The program interprets the rotation of three digital shaft encoders to mean: 1) magnify or reduce the drawing, 2) rotate the drawing clockwise or counter clockwise, 3) force or relax the perspective (by changing the position of the observer relative to the object. This causes the apparent convergence of parallel lines to change).

GRAPHICAL TRANSFORMATIONS - The four projections viewed on the scope are not four independent dig plays of stored two-dimensional information; rather, space coordinates in a single data structure are transformed into two-dimensional images for display. Rotating, translating, magnifying, and changing the perspective does not affect the data structure, (local transformations excluded). A line being drawn in any one view is simultaneously seen in the three other views; lines are in effect drawn in three-dimensions and simultaneously fed back for display. Rotation, magnification, translation, and perspective transformations are performed by a single 4 x 4 matrix (Developed by Lawrence E. Roberts at the Massachusetts Institute of Technology for use in his doctorate thesis on assembling three-dimensional descriptions of objects from their photographs). Two matrices are used for display purposes, one for the perspective view and the other for the three orthogonal views - thereby enabling the perspective view to be manipulated independently of the orthogonal views.
		The first quadrant in each subfigure shows a perspective view of the third quadrant, or front view.

3D Displays - In October, 1961, Ivan E. Sutherland demonstrated the power of the light-pen as an accurate drawing device while developing his thesis. The oscilloscope became a candidate for the display of three-dimensional information; erasure and speed problems evaporated. But how could three- dimensional information be best displayed on the two-dimensional oscilloscope screen? Perhaps direct three-dimensional presentations were needed. Two oscilloscopes could generate stereo-optic displays, but the idea was reluctantly discarded, because of the clumsiness of eye-glasses or other means of image separation.

		One industrial lab had just completed a new prototype three-dimensional display system. The principle used is simple: a translucent screen is rotated about a single axis fast enough to become "invisible." A projection oscilloscope repeatedly strobes the screen every time the surface moves into a desired position. The effect is a point of light in three-dimensional space. Generating thousands of points in this fashion developes a ghost-like image in space. The display is simple in principle, but difficult to implement because of electronic complexity. Seeing that the technology of direct three-dimensional display was inadequate, the necessity of direct display was questioned. Depth perception is aided visually by three independent means: 1) the binocular effect, 2) the perspective effect, and 3) the obstruction of rearward surfaces by opaque, forward high-lighted surfaces. The independent nature of these aids is demonstrated every time an object over thirty feet away is studied. Binocular vision fails at this distance and perspective is the next aid we unconsciously call upon. Perspective fails, also, if the object is totally unfamiliar. For example, unless previous experience has taught us that the object in question has vertical or horizontal edges and is a certain size, one cannot determine the relation of the object to its surroundings by perspective alone. If we are able to move relative to the object, or if it is moving relative to us, we call upon the third aid.
Graphical interpretation of method used in the PSL program to define a 3-D point. The cube represents the imaginary volume enclosed by the orthogonal viewing quadrants.

The two dimensional CRT screen could duplicate perspective and shading. However, shading effects due to simulated lighting and simulated surface texture are difficult to produce by computational methods, so this visualization aid was dismissed. The question then became: would perspective and surface interaction through obstruction alone provide the observer with sufficient sense of depth to permit him to visualize ordinary mechanical parts?

Although the answer is not clear even today for parts containing arbitrary surfaces, it was felt then that given the ability to move the part in perspective, so that relative motions of edges could be detected, the observer would have suitable information to establish the desired three-dimensional perception. Late in August 1962, M.I.T Lincoln Laboratory extended an invitation to use their TX-2 computer. The TX-2 computer was specifically designed for rapid input-output in the real-time domain - a pre requisite for successful graphical communication. At this time, exposure to Lawrence Robert's multiple transformation matrix occurred. His compact, unified approach to transformations sparked new thought and the present approach of drawing directly in three dimensions was born. Ivan Sutherlands Sketchpad programmed for the TX-2 computer was in operation at that time. The move to the TX-2 enabled the incorporation of many of Mr. Sutherland's utility programs into the Sketchpad III system, saving vast amounts of programming and debugging time.
		Typical graphical presentation showing top, front, and side views plus a "3/4" perspective, view.

Future - For users who are previously acquainted with projection systems, learning how to use Sketchpad III is rapid. Usually ten to twenty minutes of instruction is necessary before an understanding, although not a confident skill, is gained. People who have had no experience with projection systems cannot use Sketchpad III without first learning projection conventions.

		However, even learning projection conventions will be easier with the aid of Sketchpad III. Sketchpad III is capable of doing many things easily and well, but still falls short of general graphical communication. Research in the near future will bring the system closer to a general purpose system. Program additions that enable copies of structures to be locally manipulated will include the means of generating reflections for use in building symmetrical parts. Work is beginning on arbitrary surface manipulation, using Sketchpad III as a research tool. The future will bring exciting changes in engineering. Two-way graphical communication networks between plants and field locations will enable engineers to change designs quickly as the need arises. Several consoles in a plant time sharing one large central processor will permit several component designers to work concurrently on a system design.
The 1964 TV program with Ivan Sutherland's Sketchpad III at MIT.

With bookkeeping programs, designers would be automatically informed if a design interfered with components described in other sections of the plant. Thus fuel lines would no longer be discovered passing through servomechanisms at the mock-up stage. With system designs becoming increasingly complicated, graphical communication with digital computers is bound to lessen the increasing amount of time spent in the design stage of manufacturing.

Johnson, T. "Sketchpad III: Three Dimensional Graphical Communication with a Digital Computer," in AFIPS Spring Joint Computer Conference. 1963. 23. pp. 347-353.