Journal Note, AAS., June 13, 1996

Triumph of the Nerds--PBS TV, June 12, 1996

based on book by Robert Cringely, Accidental Empires

The story of the personal computer and operating systems industry is one of randomness, path dependence, and economies of scale.

Shared Knowledge

The early days of the industry were marked by a community of young people fascinated by a new technology which gave them great feelings of personal power. One could write instructions for this beast and it would give you something back. New knowledge was shared in a club. People delighted in solving problems and showing these solutions to the community without any thought of patenting or monetary reward. One wonders if this high tech without a huge organized lab will ever be possible again. Steven Jobs learned his trade first by inventing a blue box that could make telephone calls all around the world for free.

When IBM wanted to enter the personal computer business they discovered that the leaders in software were Gary Kildal and Bill Gates. IBM approached Kildal but were rejected because IBM appeared so demanding, e.g. no disclosure contract. Gates decided that there was a great opportunity in an alliance with IBM, no matter how one-sided. So IBM went with Gates but Gates only had programming languages and no operating system. The first operating system CPS was developed by Gary Kildal. Gates found a person (Pattison?) who had cobbled together an operating system loosely based on Kildal's. He called it MS-DOS for make shift operating system. Make-shift or not it was the missing piece that Gates needed. Its writer was paid $50,000 and did not share in future returns. Gates was paid for his continuing relationship with IBM but received no royalties from IBM. Gates did retain the right to sell MS-DOS to any other user. IBM could not imagine that anyone else could use it. IBM's reputation made the personal computer legitimate in business taking it beyond the hobbyists.

The Killer Ap

Apple sold the first assembled personal computer in 1978 containing their own proprietary operating system. They were successful, but were quickly overtaken by IBM who in 1981 brought not only reputation, but associated applications that actually made the computer valuable in business. IBM's open architecture and operating system encouraged software writers to write for it. The "killer application" was the spreadsheet which had been invented in 1979 by ?? (can't remember his name since he did not patent the idea and captured no rents. Others took the original VISICALC and elaborated it. The IBM personal calculator plus Lotus 123 were a saleable package to business while the Apple remained a fascinating toy and letter writing tool. Try applying the theory of marginal productivity here. What is the marginal product of Lotus? Or what is the MP of the PC given Lotus? Relative prices must be a matter of institutions for complements.

The basic economics of the relationship between economies of scale and path dependency is the key point here. IBM's PC and its operating system did not have to be better than Apple's. It only needed to constitute a package with software that was better. It is simply too costly once the industry is on a particular path or trajectory for a competing system to offer the products at a good price. Writers of software went with the dominant operating system which was made dominant by IBM's reputation and sense to combine itself with spreadsheets. A little initial advantage translated into a huge cost differential. People bought IBM because it had more software and this process feeds back in a non-linear fashion of cumulative causation (see my article, "Intellectual Property Rights in Bio-Technology and Computer Technology," J. of Theoretical and Institutional Economics, 141(1):127-41, 1985).

IBM's attraction of software writers as complements outside of the firm contributed to their initial success. But the same transparency eventually made them the victims of reverse engineering and the interpretation of intellectual property. Engineers made a slightly different hardware which nevertheless could run software written for the IBM. So if they could sell it for less they could undercut IBM. And Compaq and others did. But Compaq had to pay Gates royalties for the operating system. So Gates made little money from IBM, but IBM created the market for cloned PC's which used MS-DOS and did pay Gates royalties and this happenstance was the original basis for his fortune. This market demand, accumulated experience by users, and learning by doing was a high exclusion cost good. So if IBM's ROM-BIOS which connected the hardware to the operating system could be reverse engineered and copied by competitors, why were there no similar competitors for MS-DOS that could run software written for it? I don't have the technical knowledge to answer this. I am not sure if it is a matter of interpretation of intellectual property or inherent in the technology--a little different operating system does not fit the software written for the original.

Uncertainty, Vision, Strategic Planning

Xerox was a major player in office practices using paper and began to worry if new technology would destroy its markets. It set up a research facility in California which we now know was immensely inventive. But Xerox top management on the east coast never understood how it could be translated into salable products. The best staff economist trained to equate marginal revenue to marginal cost did not help Xerox management in strategic planning if the technology plus user needs was not understood. Steven Jobs did. On a trip through the Xerox facilities he saw GUEY's. Graphic interface things. He saw that this graphic way to command the computer's functions would make it user friendly beyond the demands of complicated MS-DOS commands. He went back to Apple and led the design of the Macintosh in 1984. So Apple's disadvantage in having less software written for it began to be offset by the Mac's inherently more usable graphic commands. IBM needed a GUEY too. Reverse engineering to the rescue again. Gates came up with Windows for IBM. Jobs said, "Microsoft as no taste." Apple sued but lost. What is the MP of economists designing optimum combinations of known inputs for production of the traditional products of Xerox vs Jobs who had a strategic vision connecting technology and the evolution of PC's vs. the bench engineers who invented GUEY in the first place? vs. the hackers who first made a personal computer do anything at all--and for the sheer joy of invention?

What is the value of this little story? Is it mere journalistic anecdote? Non-generalizable? There are lots of articles running regressions relating variables like education to number of patents in a country. The relationship may be scientific, but the number of patents is a poor proxy for direction setting trajectories like PC's. There is a generalizability problem in going from number of patents to specific future-shaping technologies which is not solved by large numbers and statistical tests. It takes a story.

This was written on an IBM clone running Windows! 'nuf said.

If you have any questions or comments, please email

Back to Main Page