VisiCalc — The Spreadsheet That Gave People a Reason to Buy a PC

VisiCalc was a spreadsheet program for the Apple II, created by Dan Bricklin and Bob Frankston and published by Personal Software, the company later renamed VisiCorp. It was announced publicly at the National Computer Conference in June 1979, and the first production copies shipped on 17 October 1979 — as Frankston put it in his account of writing it, "On October 17th we finally shipped the prototype, now dubbed the production version, of VisiCalc." What it did is now so familiar it hardly needs explaining: you type numbers and formulas into a grid of cells, and when you change one number, everything that depends on it recalculates at once.

Its effect on the Apple II itself — businesses buying the computer in order to run the program — is covered in the Apple II article. This page is about the program.

From a classroom daydream to a shipping product

Bricklin has described the idea forming while he sat in a Harvard classroom in spring 1978, daydreaming about a calculator with a mouse on its back and a fighter-plane heads-up display. The vision became more practical as he worked on it. A BASIC prototype on the school's timesharing system turned free placement of numbers into rows and columns with names people could use, and added the status line showing the formula behind a value. The first personal-computer prototype, which his work journal dates to 8 October 1978, was written over a weekend in Apple BASIC on an Apple II borrowed from Dan Fylstra of Personal Software, who became the publisher. The Apple II's game paddle proved too imprecise for moving the cursor, so the arrow keys took over.

The name was not Bricklin's first choice. He had proposed "Calcu-ledger"; VisiCalc came out of a meeting with Frankston and Fylstra, and Fylstra, who ran the marketing, picked it.

Written for 32KB

Frankston did the programming, at first in his attic in Arlington, Massachusetts. VisiCalc was written in assembly language for the Apple II's 6502 processor, using an assembler that ran on MIT's Multics timesharing system, which was cheapest late at night — so Frankston worked nights. The macro assembler let him write structured code with IF-THEN-ELSE-style macros.

Memory was the constraint. Frankston's goal was to fit the program, the operating system and the screen into a 16KB machine, but the code grew to about 20KB and VisiCalc ended up requiring 32KB; by the time it shipped, 48KB machines were common. The Apple II had no interrupts and no clock, and characters typed during a long calculation could be lost, so he scattered keyboard checks through the code. To avoid depending on Apple's licence terms, he also reverse-engineered the disk routines and wrote a compatible file system.

What was actually new

Bricklin is careful about the word "first". Row-and-column financial programs existed before VisiCalc, he writes, and Frankston had worked at a timesharing company that ran some of them. What he lists as special is the combination:

• formulas stored in cells that recalculate instantly and automatically when the cells they reference change
• input, definition, formatting and output merged into one program-by-example interface — you build the model by doing the calculation once
• replication of any range to any other, with relative and absolute references
• all of it on an affordable personal machine, where earlier forecasting tools ran on timesharing systems that could cost more than $1,000 a month

These traits — labels and formulas distinguished by the first character typed, the A1-style grid, copying with relative references — survived into the spreadsheets that followed.

Why it was never patented

VisiCalc was protected by copyright and trademark, not patent. Bricklin explains that in 1979 software inventions were rarely patented, since programs were widely treated as mathematical algorithms. The patent attorney Personal Software retained estimated about a 10% chance of success, and given that and the cost, they did not file. In 1981 the US Supreme Court ruled that a claim did not become unpatentable simply because it used a computer program, opening the way to software patents — too late, Bricklin notes, for the spreadsheet.

The killer app

In a 1996 Japanese television documentary that Bricklin has posted, Steve Jobs recalled Dan Fylstra arriving at his office with a disk and a program he called a visual calculator, and said "that's what really drove -- propelled -- the Apple II to the success it achieved." (Jobs placed it in 1977; Bricklin points out that it was 1979.)

VisiCalc soon left the Apple II. Frankston describes versions for the Commodore PET and the Atari 800, which used the same processor, and a version was among the programs listed as available for the IBM PC when IBM announced it in 1981. Personal Software renamed itself VisiCorp in 1982.

On the IBM PC, VisiCalc was overtaken by Lotus 1-2-3, released in January 1983 and written to take full advantage of the new machine. In 1985 Lotus acquired Software Arts.

See also the hardware timeline, the overview timeline and Apple.