DAC-1 (Design Augmented by Computers) was the first computer-aided drawing system, created by Don Hart and Ed Jacks at General Motors and IBM. The system originally was based on an IBM 7090 computer (later upgraded to a 7094 in 1963), augmented with extra disc space and a specially designed IBM 7969 "image-processing system." Input was with punch cards, but it also was capable of scanning in drawings. It allowed the user to input a 3D description of an automobile and then rotate it and view it from different directions. The final data could be output to either 35mm film (by way of a CRT), a hard copy plotter, or used to drive computer-controlled machining devices. It was unveiled at the fall Joint Computer Conference in Detroit 1964.

In 1959, Donald Hart encouraged the applications group to begin the implementation of such a system, and a five-person project team was formed. The project was named "Digital Design," and it went full steam ahead to develop a feasibility demonstration. The name of the project was later changed to DAC-1, Design Augmented by Computers, since GMR management did not want to give the impression that digits were being designed.

		Using the IBM 704 & IBM 780 display unit for an interactive input device - What resources were needed to carry out a feasibility demonstration? Members of the department had already programmed the IBM 704 computer to produce a point plot on the IBM 780 display unit. To maintain an image on the display only required putting the program into a loop. Interactive control of the computer was solved in a unique fashion. The normal control of the system was via sense lights and sense switches on the faceplate of the central processing unit. Programs were started by keying in bootstrap commands through these sense switches. Clearly, this was inadequate for any type of interactive design activity. Attached to the central processing unit was a line printer that was normally used for operator messages. An arrangement of five 10-position dials was wired into the echo checking logic of the on-line printer. A monitor program was then written that would poll the printer and read the positions of the five dials.
1956 Lawrence Livermore National Labs uses a IBM 740/780 to create recordings of B&W Motion Pictures.

These five digits were then used to control program execution, thus providing the project with an interactive input device. With this arrangement, it was now possible to control the software and to direct program execution of the IBM 704 computer interactively.

The display unit and numerical control machines provided good methods for the creation of graphic output and physical models. The same boundary curve and section line drawings that were viewed on the display screen could also be redirected to a GMR-designed drawing table that was mounted on a three-axis milling machine. Milling Machine used for the DAC-1 - The spindle of the milling machine was simply used to hold a ballpoint pen. This idea had already been used some time earlier at Boeing Aircraft to produce full-size drawings or "lofts" of airfoils. The only added flourish was to divide the full-size drawing into 32x32-inch segments that could be drawn one segment at a time. The drawing table was then equipped with an indexing head and set of rollers that increased the effective drawing area to 96 inches by the length of the length of the roll.
		The DAC-1 System at GM Research Labs, 1965.

The Photomultiplier Tube - There still remained the problem of providing graphical input to a computer and, with the help of the GMR Instrumentation Department, this aspect of the problem was addressed at the same time. A digital circuit was designed to use the signal from a photomultiplier tube mounted in a hood covering the display unit to detect the presence or absence of light and set one of the sense lights on the IBM 704 computer accordingly. Drawings and sketches were traced onto a clear plastic overlay mounded on the face of the IBM 780 display unit. The lines on the overlay were then digitized by finding those points where the light, from a spot of light plotted on the display unit, was occluded by the line on the overlay. Since a drawing consisted mainly of white space and the actual lines themselves covered only a minute percentage of the total area, digitizing was not done by a raster scan. This would have been much too slow using existing scan rates. Instead, program logic was developed to "lock onto" a line and digitize the path to each terminal point in much the same manner as an ant might crawl along a wire.

The Digital Design demonstration had to show how computers could be applied to the vehicle design process. Also, in August 1959, preliminary discussions were started with IBM for the acquisition of custom-designed graphics hardware and an IBM 7090 computer. Therefore, the demonstration had to do more than demonstrate feasibility. It had to justify the investment in a multimillion-dollar joint project with IBM for the development of a laboratory for the study of graphical man-machine communication.

The feasibility demonstration was designed to show how a mathematical model of a vehicle might be created starting from a series of designer sketches. After the three-dimensional mathematical model is created, it must be capable of being modified quickly and then the resulting mathematical model used to create physical models and full-size drawings. All the mathematical algorithms were to be programmed for the IBM 704 computer using the new Fortran language and compiler. System routines or I/O drivers were coded in assembly language.

During discussions with Styling, Fisher Body, and other divisional personnel, it became apparent that they were very concerned about the smoothness of the initial line drawings used to show a new styling concept. Line drawings must represent aesthetically pleasing curves from which all irregularities, flat spots, or hollows have been removed. The conventional process was to use the skills of a trained clay modeller to translate from a rough design sketch to a full-size clay model of the vehicle. It was felt that a design machine could address some of these same problems if it were introduced early enough in the design process. Therefore, it was decided that early styling concept sketches would be the starting point. Stylists were requested to sketch the concept in perspective on paper, and the computer would then be used to construct a three-dimensional mathematical model that not only reflected the design intent but also rep-resented the object to the necessary degree of detail, smoothness, and accuracy. A transparent overlay was used to trace the location of four boundary curves of the hood surface. The two-dimensional shape of these boundary curves, along with the orientation and position of the model in Cartesian coordinate space, provided enough information from which to compute the spatial location of the hood boundary curves.