Riddle’s account comes primarily from his own writings and an extensive interview he gave historian David Weisberg in April 2001 — nearly two decades after the events, and after a bitter legal dispute ending in an $11.875 million settlement. Weisberg’s history draws on Riddle’s interview alongside Walker’s Autodesk File and other primary sources. Walker’s Autodesk File — 900 pages of internal documents compiled by Walker himself — is the other major primary source. The two accounts align on most facts and diverge sharply on questions of credit and intention.

What follows draws on all of it. Both accounts have their perspective. The facts both sides agree on are the foundation. Where they diverge, both versions are noted.

Riddle was born in California and built his first computer in junior high school out of relays. After Arizona State University he went to work for a steel fabricator doing structural steel detailing for the Palo Verde nuclear power plant west of Phoenix. The company used a ComputerVision CADDS3 system — a $250,000 professional CAD workstation — for strictly 2D drafting work. Riddle studied it and concluded a microcomputer could do everything they were doing on that $250,000 machine. He was right, though it would take him years to prove it.

Around the same time he began working at a local ComputerLand store, which gave him free after-hours computer time. He used it to start writing Interact. He was also consulting for the Frank Lloyd Wright Foundation in Scottsdale, watching architects actually design. Most engineering software of the era was built by engineers who had never sat with a designer. Riddle had. Interact reflected it.

Interact took two years to reach a shippable state — not because Riddle was slow but because the hardware kept moving. He was writing in pieces, assembling the program as larger memory boards became available. The fundamental problem was arithmetic: CAD requires continuous floating-point calculations, and early microprocessors did floating-point in software, which was prohibitively slow. Riddle concluded he needed a processor with hardware multiply. This requirement is what led him to Walker.

Walker had founded Marinchip Systems in California in 1977, building computers around the Texas Instruments TMS-9900 — a 16-bit chip TI had never properly marketed, but which had hardware multiply. Riddle was already doing freelance consulting work for Walker and Drake, producing utility software Marinchip bundled with their systems. That existing relationship was how he encountered the hardware Interact needed. When Walker saw Interact running on a Marinchip machine, he became a dealer for the software. The two men were connected not by the UNIVAC network that assembled most of the other founders, but by hardware necessity.

Interact shipped in 1979. Its first commercial customer was Atlantic Richfield — ARCO — which used it to plan deep dives for offshore oil drilling rigs. The setup comprised an S-100 machine with dual 8-inch floppy drives, a 640×480 graphics board, a Houston Instruments digitizer, a Televideo terminal, and a Rapidograph pen plotter. A complete professional drafting workstation at a fraction of the price of anything comparable.

Interact was written in a proprietary programming language Riddle had invented. The compiler for that language was written in another language he had also invented. Greg Lutz later described this as extreme NIH — Not Invented Here — syndrome. It is a fair characterization and also an incomplete one. In 1977, the available microcomputer compilers were slow and generated inefficient code. Riddle’s decision to write his own toolchain was a pragmatic response to what existed. The problem was what it meant for anyone else trying to work with the code later.

When Autodesk needed to port Interact to the IBM PC, Drake and Lutz had to read code in a language documented nowhere outside Riddle’s own head. They split the source files alphabetically — Drake from A, Lutz from Z — and converted to C without consulting Riddle. This is documented in the 2020 CHM founders oral history. By the time AutoCAD 1.0 shipped in December 1982, the code running it was almost entirely new. Walker’s own trivia quiz records the number of Riddle’s original lines that survived: fewer than 1,200 out of 12,000 total.

The number requires context. What survived wasn’t just code — it was architecture. Specifically the E struct: Riddle’s unified entity data structure, one instance per drawing object, that defined how AutoCAD represented every geometric element. That data model carried through the C rewrite intact because it was right. The lines of code were Autodesk’s. The design those lines expressed was Riddle’s.

In late 1981, Walker assembled eighteen people at his house in Mill Valley. Riddle was among them. The deal: a non-exclusive license for Interact in exchange for a ten percent royalty on all AutoCAD revenue and any derivative products, with Riddle retaining the right to develop competing software independently.

According to Riddle’s account in the Weisberg interview, the non-exclusive terms were Walker’s idea — Walker liked contractual simplicity and cited a page-and-a-half Colgate Palmolive deal as his model. He almost certainly underestimated what the royalties would eventually amount to. When Autodesk was selling a few hundred copies of a $1,000 program, ten percent was manageable. When it was selling tens of thousands of copies annually, ten percent was a significant obligation the company had no clean way to exit.

Some context behind the founding comes from something Walker told me personally during a visit to Neuchâtel, Switzerland for the 25th anniversary of AutoCAD.

When Walker heard that new 8086 personal computer systems from IBM were coming to the US market, he asked what software had been developed for them. He was told: none that anyone knew of. So he set out to have several programs ready for the impending new machines. That’s the logic behind the five-project strategy — a deliberate attempt to have software on the shelf the moment the IBM PC created a market. Interact was one of the five. It happened to be the one that hit.

Walker also described what happened at the 1982 West Coast Computer Faire. The CAD demonstration got so much attention that they found themselves suddenly in the business of CAD. The crowd’s first request was for dimensioning. Nobody on the team had anticipated that. Walker’s Christmas 1982 hack — writing the entire dimensioning package in three days — was the direct result of what they heard at that booth.

Both accounts agree on the facts of what happened next. Interact in 1979 was a remarkable achievement. It was also running on specialized hardware in a proprietary compiled language, sold in tiny quantities to a specialized market. That is not a path to the global standard for CAD software used in 80 countries in 18 languages.

The IBM PC port. Drake and Lutz’s conversion to C made AutoCAD compilable for any platform with a C compiler. Interact’s proprietary language locked it to whatever hardware Riddle had built his compiler for. The C rewrite was a platform-independence decision that shaped the product’s entire market trajectory.

The open architecture. Walker’s insistence on open, extensible design — publishing DXF, developing AutoLISP, inviting third-party developers — created the ecosystem that made AutoCAD a standard rather than a product. Riddle built a drafting program. Autodesk built a platform.

The distribution and sales infrastructure. Mike Ford, Autodesk’s first marketing hire, is credited by Riddle himself in the Weisberg interview as essential to the company’s success. The dealer network, trade show presence, and Comdex demonstrations in sixteen simultaneous booths were not Riddle’s work.

The development velocity. Walker’s holiday hacks — the dimensioning package in three days, AutoLISP entity access, the Urgent Fury UI overhaul, the solid modeling prototype in one day — added capabilities Interact had never had. The product by Release 9 was almost unrecognizably different from what Interact had been.

The prospectus describes the C implementation and open architecture as the product’s distinguishing technical features. It does not mention Riddle or Interact. From Autodesk’s perspective by 1985, AutoCAD’s value was in what Autodesk had built around the original architecture. From Riddle’s perspective, that architecture was what made everything else possible. Both perspectives have merit and neither is complete.

The royalty agreement became a source of tension not because either party acted in bad faith at the outset but because the original terms didn’t anticipate the scale of what AutoCAD would become. By fiscal year 1984, Riddle was receiving nearly $600,000 a year from a program that had been almost entirely rewritten. By the late 1980s that figure exceeded $1 million annually. Autodesk treated major AutoCAD extensions as separate products — ADE-1, ADE-2 — that fell outside the royalty base. Legally defensible. Also exactly the kind of behavior that poisons a business relationship.

Walker’s trivia quiz notes that while still being paid by Autodesk, Riddle developed a competitive product distributed by VersaCAD Corporation. Worth clarifying: VersaCAD and EasyCAD were not the same product. VersaCAD was T&W Systems’ own CAD program. In April 1986, VersaCAD Corporation began distributing Riddle’s EasyCAD through its channel at $495. Riddle’s original license agreement explicitly gave him the right to develop competing software. Walker put the trivia question in anyway. Both things are true and both were resented.

Riddle’s version: Autodesk weaponized the agreement’s ambiguity against him. Autodesk’s version: the non-exclusive license created obligations Riddle wasn’t honoring. After years of mounting hostility, Riddle and his wife filed suit in August 1991. Settlement, January 1992: $11,875,000.

Riddle founded Evolution Computing and produced EasyCAD (1985), FastCAD (1987), and later FelixCAD. EasyCAD was among the first CAD programs under $500, bundled with the Microsoft Mouse for three years. FastCAD was the first CAD program written entirely in x86 assembly language. None of these products achieved anything approaching AutoCAD’s market position. FelixCAD was acquired by a German company and eventually folded into other products.

The CAD Society gave Riddle its Lifetime Achievement Award in 2008. Engineering News-Record named him one of the top 125 people in the construction industry over 125 years. Design News nominated him for Engineer of the Year in 1992. Meaningful recognitions for a genuine historical contribution.

Walker’s trivia quiz, Question 49: “Why did the police seize a workstation containing AutoCAD source code?” Answer: “The Apollo workstation in John Draper’s apartment was seized when his roommate was arrested for allegedly counterfeiting BART tickets.”

John Draper is Captain Crunch — the phone phreaker who cracked AT&T’s long-distance network with a cereal box whistle, who influenced Wozniak and Jobs before they founded Apple, who wrote the IBM PC’s first word processor from a jail cell. He worked at both Apple and Autodesk. An Apollo workstation in his apartment — being used for Autodesk development — was seized when his roommate was arrested for counterfeiting Bay Area transit tickets. The AutoCAD source code went with it. Documented in Walker’s own words.

Mike Riddle built something real. He saw that a $250,000 ComputerVision machine was doing work a microcomputer could do, and he built the microcomputer version before almost anyone else thought to try. The architecture he created — particularly the entity data model that survived the C rewrite — was sound enough that Drake and Lutz kept it. The CAD industry today descends directly from decisions Riddle made alone in Arizona in 1977 and 1978.

It is also true that Interact, in its original form, would almost certainly not have become the global standard for computer-aided design. Written in a language nobody else could compile, running on hardware nobody else was selling, with one customer. The rewrite to C, the open architecture, the dealer network, the open DXF format, AutoLISP, the Comdex launch — those were Autodesk’s work. AutoCAD became AutoCAD because of both.

The $11.875 million settlement measures the legal value of an original contribution to a product that had largely moved past it. Both sides decided that number was close enough to avoid finding out what a court would award.

Riddle’s blog, written years after all of it, is worth reading for what it says about him as a programmer. His posts on debugging, floating-point arithmetic, and program design are the writing of someone who remained genuinely engaged with hard problems across decades. The same instincts that produced Interact in 1977 were still active in 2009. The CAD industry he helped start went in a direction AutoCAD defined. His subsequent work explored the architecture questions that interested him most, on his own terms.

Primary Sources

CHM Oral History #1
“AutoDesk Oral History Panel: Dan Drake, David Kalish, Duff Kurland, Greg Lutz”
Interviewed by Chuck House and Burt Grad via Zoom · October 15, 2020 · 49 pp.
CHM Reference: X9368.2021 · Accession: 102792114-05-01 · © 2020 Computer History Museum

The Autodesk File, Fifth Edition (2017)
John Walker · fourmilab.ch/autofile
Chapters cited: AutoCAD-80 Development Log; IPO prospectus; Autodesk Trivia Quiz (chapter2_112) — Q.12, Q.49, Q.52. Freely available online.

Additional Sources

Weisberg, David. “Autodesk and AutoCAD.” History of CAD. shapr3d.com/history-of-cad/autodesk-and-autocad
Based on interview with Mike Riddle, April 28, 2001.

Riddle, Mike. michaelriddle.com — Personal website and blog.

DigiBarn Computer Museum — “Mike Riddle & the Story of Interact, AutoCAD, EasyCAD, FastCAD & more”
digibarn.com/stories/mike-riddle/

Wired, April 2007 — Profile of Mike Riddle. wired.com/2007/04/mike-riddle-cad/

Wikipedia: “AutoCAD”; “Autodesk”; “FastCAD”; “John Walker (programmer)”