AutoCAD 2.1 shipped Lisp as a "variables and expressions feature" because Autodesk feared the word would scare developers

Autodesk had a full Lisp interpreter and a marketing problem with the first four letters of its name.

The company’s own account is blunt. In the preface to a position paper on its site history, the team writes: “when we first shipped AutoCAD 2.1, we didn’t use the word “Lisp” at all.” What they called it instead was the “variables and expressions feature.” It was the sort of label you’d expect on a spreadsheet’s third menu.

The reason was the calendar. This was the middle of the expert-systems hype, when Lisp sounded like something that lived inside an AI lab, and the worry was that it sounded like arcane knowledge. A CAD user who just wanted a macro did not need to be told they were about to learn the language of artificial intelligence.

So the plan was to ship the capability first and the name later. Per the same preface, “Only in release 2.18, in which we provided the full functional and iterative capabilities of Lisp, did we introduce the term “AutoLisp”.” The disguise came off once the thing could do the full job.

The paper itself is by John Walker, Autodesk’s founder, dated February 5, 1985, and it reads like someone talking himself out of the embarrassment in advance. Walker is not shy about the reputation. He calls Lisp “the most arcane, obscure, and hopelessly-rooted-in-the-computer-science-department language in the world” before arguing that it’s the right tool anyway.

His case has three legs. The first is pedigree: “Lisp is the preeminent language in the field of Artificial Intelligence, and has been for over two decades.” He backs it with a system called NAVEX, which he says is “written entirely in Lisp” and “will soon be ensuring that the Space Shuttle reaches the runway.” (If your macro language is good enough for orbiters, your hatch pattern probably survives.)

The second leg is how designers work. C and FORTRAN, Walker writes, “force one to organise a problem entirely before programming,” while “Lisp encourages exploring various approaches to a problem interactively, exactly as CAD helps a designer.” That is a fair description of drawing: you move a line, look at it, and move it again.

The third leg is humility. Walker is clear about where Lisp stops: “Lisp is the language we sell for small applications–we are not offering it or suggesting it for major programming projects.” Bigger jobs, he says, were for other interfaces. So the language that had to be disguised was also being sold with a ceiling on it.

The pattern is old and it will keep working. Developers distrust a new language until it has already solved their problem, and then they claim they liked it all along. Anything that arrives labeled “scripting” or “expressions” tends to be judged by what it does, not what it’s called.

It also lines up with today’s habit of selling the same technology under three different brand names depending on who is in the room. Autodesk did it with one word in 1985, and at least did it to avoid scaring people, not to inflate a price.

Somewhere in the middle of all that careful reasoning, Walker also writes two words that aren’t an argument at all: “So there!”