Warren Robinett

Warren Robinett is an American designer of interactive computer graphics software and computing hardware who worked on early virtual reality systems at NASA Ames Research Center and the University of North Carolina at Chapel Hill (UNC). In the mid-1980s he developed the software for the first prototype of NASA's Virtual Environment Workstation, including the glove-based gesture interface that let users point to fly through a virtual world. At UNC he managed the Head-mounted display Project from 1989 to 1992, co-wrote papers on the optics and display mathematics of head-mounted displays, and with R. Stanley Williams conceived the nanoManipulator, a virtual reality and force-feedback interface to a scanning tunneling microscope.

Before his VR work, Robinett designed the Adventure cartridge for the Atari 2600, which contained a hidden room displaying his name, and co-founded the educational software publisher The Learning Company. In the mid-1990s he ran Virtual Reality Games, Inc., a Chapel Hill start-up that built a prototype game for Hasbro's unreleased home VR system, and he later spent 2003 to 2012 at HP Labs working on defect-tolerant computing and memristor circuits.

Reviewed 4 October 2026. Career dates, NASA Ames, UNC, nanoManipulator, Virtual Reality Games and HP Labs facts checked against Robinett's CV and site, Fisher 2016, Taylor 1994, the 1991 Whole Earth Review scan, the NRC roster, Crossref metadata for all cited papers, the Halcyon Days interview, GDC Vault, SIGGRAPH History and SPS pages. About review dates.

Early life and education

Robinett graduated from Glendale High School in Springfield, Missouri, in 1970; Springfield Public Schools inducted him into its SPS Hall of Fame on October 16, 2025. He earned a B.A. in "Computer Applications to Language and Art" from Rice University in 1974 and an M.S. in computer science from the University of California, Berkeley, in 1976.

Career

Atari and The Learning Company

Robinett started work at Atari as an Atari 2600 game designer in November 1977. While working on Slot Racers, he played the original text adventure game by Will Crowther and Don Woods at the Stanford Artificial Intelligence Lab in 1978. He then began a graphical version for the console, which he completed in June 1979 alongside the Basic Programming cartridge. The game fit 30 rooms and 18 objects into a 4 KB ROM. Because Atari did not credit its designers on packaging, Robinett hid a secret room that displayed his name and could be entered only with a hidden object he called the Gray Dot; he told interviewer James Hague that the first finder he knew of was a fifteen-year-old from Salt Lake City who wrote to Atari explaining how to get in. Robinett's CV states that Adventure sold one million copies. He left Atari in June 1979.

In 1980 Robinett co-founded The Learning Company with three educators and served as its chief software engineer until 1983. With Leslie Grimm he created Rocky's Boots, an Apple II and IBM PC game in which children built simple digital logic circuits on screen, as well as Gertrude's Puzzles and Gertrude's Secrets. Rocky's Boots was shown at SIGGRAPH 1984.

NASA Ames Research Center

According to his CV, Robinett was a research scientist on the Virtual Environment Workstation project from 1986 to 1987, based at NASA Ames Research Center in Mountain View, California. Scott Fisher, who directed the project, has written that Robinett was one of the people he eventually recruited to help build the system, together with Mark Bolas, Scott Foster and Steve Bryson, and that he negotiated a contract with VPL Research, the company's first, to build the first "dataglove" for the project.

Robinett's site states that he developed the software for the first prototype in collaboration with NASA scientists Fisher and Michael McGreevy, and describes that prototype as a motorcycle helmet fitted with Watchman LCD displays, wide-angle stereoscopic optics and a Polhemus magnetic head-position tracker. The system's glove measured the bending of the finger joints and carried a tracker for the hand's position and orientation. He designed the display and interaction software and the glove's gestural command set: pointing with the index finger flew the user through the virtual world, grabbing moved virtual objects, and other gestures scaled the surrounding world up or down. He credits the system with the first use of a glove in VR. His demonstrations included a computational fluid dynamics flow field the user could stand inside, a simulated escalator that carried the user through the world, a model of the proposed space station, and a molecule that could be scaled up and down. The system was described in "Virtual environment display system" by Fisher, McGreevy, James Humphries and Robinett, published in the proceedings of the 1986 Workshop on Interactive 3D Graphics.

University of North Carolina head-mounted display research

Robinett joined the UNC Department of Computer Science in 1989 as a research associate and manager of its Head-Mounted Display Project, directing a team that included six graduate students, an optical engineer and a laser physicist. The group explored applications including architectural walk-throughs, molecular graphics, medical imaging and 3D modeling. With Richard Holloway he showed "Flying Through Protein Molecules" at SIGGRAPH 1991, and he presented at the SIGGRAPH courses "Implementation of Immersive Virtual Environments" (1992) and "Implementing Virtual Reality" (1993).

Papers from this period addressed how a VR system should compute and display its images:

• "A Computational Model for the Stereoscopic Optics of a Head-Mounted Display" (1992), written with Jannick P. Rolland for the first issue of Presence: Teleoperators and Virtual Environments, defined orthostereoscopy in VR as the constancy of the perceived size, shape and position of simulated objects as the head moves. It argued that this requires the graphics model to match the HMD's physical geometry, named ignoring optical distortion and ignoring variation in interpupillary distance as two common modeling errors, and calculated the model's parameters for the VPL EyePhone.
• "Implementation of Flying, Scaling and Grabbing in Virtual Worlds" (1992), with Holloway, at the ACM Symposium on Interactive 3D Graphics.
• "Synthetic Experience: A Proposed Taxonomy" (1992) classified technologically mediated experience, from the microscope and telephone to VR and teleoperation, as a sensor-to-display link from the world to a person and an action-to-actuator link back.
• "Scanned Laser Displays for Virtual Reality: A Feasibility Study" (1993), with Douglas Holmgren of UNC's Department of Physics and Astronomy, reviewed displays that raster-scan a laser onto the retina. It judged a monochromatic laser-diode system with acousto-optic and galvanometer scanners the most practical choice for an HMD at the time, estimated that 1000 by 1000 pixels at 60 frames per second should be possible, and concluded that a full-color version could not yet be built from off-the-shelf parts.
• "The Visual Display Transformation for Virtual Reality" (1995), with Holloway, set out the full chain of transformations from object coordinates to the screen coordinates of each HMD display, covering head tracking, adjustable eye separation, off-center projection and lens distortion correction, as implemented in the UNC VR software. It also introduced a vector-quaternion-scalar (VQS) representation for transformations between coordinate systems.

In "Electronic Expansion of Human Perception", published in the Fall 1991 issue of Whole Earth Review, Robinett proposed linking electronic sensors to head-mounted displays. One example was a see-through HMD that optically superimposed an ultrasound scan, so that an obstetrician would see the image of a fetus appearing inside the mother's body; he called this "X-ray vision". The article also described a tracked see-through HMD that registered spatial data files with the real world, such as labels attached to objects or a route marker hovering above a road, and credited media technologist Michael Naimark with the term "real-space imaging". Robinett's own site describes the article as one of the first descriptions of what is now called augmented reality and notes that Tom Caudell coined that term the following year.

NanoManipulator

According to Robinett, he and R. Stanley Williams conceived the nanoManipulator in 1991, and UNC graduate student Russell M. Taylor II implemented it as his PhD project. Robinett was co-director of the project and a co-principal investigator on its National Science Foundation funding. Taylor's 1994 dissertation, supervised by Fred Brooks, thanks Robinett "for getting me involved in this project". It records that in January 1992 a scanning tunneling microscope (STM) built at UCLA was brought to North Carolina and connected to the hardware and software of UNC's head-mounted display and GROPE force-display projects. The user interface combined a stereoscopic color head-mounted display, a force-feedback remote manipulator and a high-performance graphics computer. Users could see and feel the sample's surface and control the timing and location of voltage pulses applied between the STM tip and the sample.

The system was presented at SIGGRAPH 93 in "The nanomanipulator: a virtual-reality interface for a scanning tunneling microscope" by Taylor, Robinett, Vernon L. Chi, Frederick P. Brooks Jr., William V. Wright, R. Stanley Williams and Erik J. Snyder. Robinett's site states that the system later used an atomic force microscope, with graphics and haptics displays, to let users push individual viruses, strands of DNA and carbon nanotubes, and that the project grew to involve faculty and students from seven UNC departments.

Virtual Reality Games, Inc.

From 1993 to 1997 Robinett was founder and chief executive of Virtual Reality Games, Inc. in Chapel Hill, a company formed to make VR games for the home market. According to Robinett, his game design document won a US$500,000 contract from Hasbro, and the company produced Siege Perilous, a prototype medieval adventure game for Hasbro's planned consumer VR system. Hasbro's home VR project was discontinued in 1996 without a product reaching the market. Robinett wrote that the 1995-era hardware was not adequate for the game and that he closed the company when a consumer market for VR entertainment did not develop, after paying its debts.

Research committees and editing

Robinett served from 1991 to 1994 on the National Research Council's Committee on Virtual Reality Research and Development, which recommended a national research agenda for the field. The committee's 1995 report, Virtual Reality: Scientific and Technological Challenges, edited by Nathaniel I. Durlach and Anne S. Mavor, lists him as a member affiliated with Virtual Reality Games, Inc. He was an associate editor of Presence: Teleoperators and Virtual Environments from 1991 to 2000 and an invited speaker at the International Conference on Artificial Reality and Tele-existence (ICAT) in 1992 and 1998. His ICAT '98 paper proposed switching a teleoperator system among teleoperation, simulation, replay and robot modes, using the nanoManipulator as the example.

Later work

Robinett returned to UNC as a research associate and research associate professor from 1998 to 2000, rejoining the nanoManipulator team. From 2003 to 2012 he was a contractor in R. Stanley Williams's Quantum Sciences Lab at HP Labs, Hewlett-Packard's research laboratory in Palo Alto, researching defect-tolerant computing, memristor crossbar memories and memristor-based logic circuits; his CV lists eight US patents from that work. Publications from this period include "Computing with a trillion crummy components" in Communications of the ACM (2007) and "A memristor-based nonvolatile latch circuit" in Nanotechnology (2010). He gave a classic game postmortem on Adventure at GDC 2015, and his CV states that between 2013 and 2016 he wrote The Annotated Adventure, a book about the game's implementation.

See also

• History of virtual reality
• DataGlove
• Head-mounted display
• Telepresence
• Ivan Sutherland