Reading the 2002 edited volume, From 0 to 1: An Authoritative History of Modern Computing, I came across an interesting history of the first software patent—a business history, as opposed to a legal history. I hadn’t seen this anywhere before, so I’ll recount it here.
Luanne Johnson, president (now co-chair) of the Software History Center, tells the story of Martin A. Goetz at Applied Data Research (ADR), a Princeton, New Jersey company founded in 1959 to sell computer programming services.
In 1964, computer manufacturer RCA approached ADR about writing a flowcharting program that RCA would provide to users of its RCA 501 computer at no cost. ADR designed and wrote the program, AUTOFLOW, and offered it to RCA for $25,000. But RCA didn’t want it at that price. Marty Goetz then went to work on a different approach to recouping the $10,000 his company had laid out to write AUTOFLOW.
There were only hundreds of companies using the RCA 501, to whom he might have sold directly. So, seeing a larger market among users of the IBM 1401, Goetz and his colleagues re-wrote AUTOFLOW for that computer. They ultimately produced superior flowcharting software to what IBM offered its customers. AUTOFLOW was capable of flowcharting the logical sequence of existing software, easing the design of software to compliment what was already in use on IBM machines. Writes Johnson:
AUTOFLOW sold quite well. As a matter of fact, considering that in 1965 no one had ever sold a significant number of copies of a software product for a price, it sold remarkably well. But what should have been a natural market for thousands of potential customers was severely constrained by the belief on the part of those customers that the product that ADR was selling for $2,400 could be gotten for free from IBM.
* * *
Over and over again, as a result of a sales presentation on AUTOFLOW, the potential customer would call his IBM account representative and ask when they were going to upgrade their program to do what AUTOFLOW did. . . . [T]he more aggressively that ADR sold AUTOFLOW, the greater the likelihood became that IBM would produce a product with comparable functionality and offer it for free, effectively putting ADR out of the software products business.
Goetz took defensive action. He applied for a patent on AUTOFLOW and served notice on IBM that they might be violating ADR’s patent application if they produced an automatic flowcharting program.
Goetz’s ADR became the first recipient of a patent on software in 1968, “a clear turning-point in the recognition of software as a product, not a service.” With attorney Mort Jacobs, Goetz determined that their software should be licensed rather than sold to customers to prevent them sharing programs, “a customary and accepted practice” at the time.
This anecdote doesn’t prove anything, but it offers a lot of food for thought and discussion. To wit:
- Johnson’s account favors today’s software patenting status quo. She refers to the “natural market” for ADR’s software, implying that IBM’s practice of writing and giving away software would have been an unnatural influence. She portrays Goetz’ pursuit of a patent as “defensive,” though there’s an equally plausible interpretation in which IBM plays defense against an insurgent by writing software to match its offering. Blocking that option by seeking a patent would part of going “on offense.” Finally, Johnson characterizes the ADR patent as “recognition” of what she appears to take as the right outcome.
- ADR’s 1968 patent on AUTOFLOW was a close predecessor to the 1969 antitrust suit against IBM, which resulted in the unbundling of hardware, customer training, system engineering, and software. Had these things not happened, the separation of the computer business from software, which we take for granted today, may not have happened as quickly. It would be a mistake to assume that software would not have emerged as a separate business line, though. The personal computer grew from very different roots than the IBM computers of the late 60s and early 70s. I think there would still be a software market separate from computers today even if the software patent and the federal government’s antitrust action against IBM had not happened. Software-writing insurgents would have had to work that much harder to make inroads, and computer companies that much harder to keep them out—all to the benefit of consumers.
- I was most interested to see that the first use of software patenting was to change the balance of power in the marketplace for existing software. The reward offered by patenting is not what caused ADR to produce AUTOFLOW, which would have been consistent with the theoretical foundations of intellectual property law. Goetz and ADR used patent to gain competitive advantage on a larger, better established rival in the market for an existing product. We may sympathize with the little guy, of course—David to IBM’s Goliath—but established firms today almost certainly make more use of software patenting as a tool of competitive advantage than upstart start-ups. Rebalancing the market for existing products, of course, is not what intellectual property law is supposed to do.
You can comment or blog elsewhere about this story or my observations on it. I don’t spend a lot of time on software patents, but I do like to collect evidence about that policy when it makes itself available.