Notes on Xerox PARC
“Xerox could have owned the entire computer industry, could have been the IBM of the nineties, could have been the Microsoft of the nineties." - Steve Jobs
I'm doing research into organizations that successfully drove breakthroughs in both basic and applied science in order to determine if there are any useful lessons to be learned for modern systems scientists.
Last week I shared my notes on Bell Labs and today I'm sharing my insights from reading about Xerox PARC.
Freedom of Research
PARC was founded in 1970 as a specialized research division of Xerox, which had established itself as a dominant force in the photocopying market.
Xerox’s CEO C. Peter McColough issued a call to turn the organization into the “architect of information” for the business office, and PARC’s founding research goal was to build that architecture. Throughout the 70s, researchers at PARC went on to invent many of the technologies that propelled humanity into the information age.
From developing Ethernet and LAN for standardized communication between computers, to introducing viable personal desktop computers, PARC’s researchers accomplished incredible feats guided by their high-level shared vision of “creating the office of the future.”
PARC’s founding research director, George Pake, held a set of core principles about managing research that were clearly conducive to the process of supporting creative invention.1
A few key ones included:
Recruit the best and most creative researchers
Don't command researchers to invent desired outcomes — describe general goals and solicit proposals
Don’t be tight with spending — give researchers the most supportive environment including plenty of the most advanced technical instruments
Research scientists enjoyed tremendous freedom and there were no standard rigid hours. Projects weren’t assigned from above — researchers generated their own ideas and formed groups based on shared interest.
Lynn Conway, who helped facilitate a revolution in the process of computer chip design while at PARC, felt that the state of art computer networking technology researchers enjoyed endowed them with "powerful invisible weapons." Their environment empowered them “to outfox and outmaneuver people who would think [they] were crazy.”
Systems Builders in Action
"Systems research requires building systems" - Jim Mitchell (PARC computer science lab)
PARC started out with three separate labs: computer science, general science, and — to my incredible surprise — systems science. I had never heard of an organization, let alone a giant like Xerox having a dedicated systems science division.
Reflecting on his experience at PARC, Pake notes the extreme rarity of systems science departments in universities and even asserts "there really is not a discipline named systems science."
And yet, PARC’s systems science laboratory, in the absence of any solid theoretical foundations for the science, managed to play a pivotal role in bringing a futuristic vision of office computing to life.
Key inventions facilitated by the Systems Science Lab included: laser printers, the object-oriented programming (OOP) paradigm, ethernet, and LAN.
The lab also helped create the Alto personal workstation, a pioneering system which introduced features like graphical user interfaces (GUIs) with overlapping windows and clickable icons.
Another remarkable example of applied systems science at PARC was the work that Lynn Conway and Carver Mead did in creating a systematic approach to the design of integrated circuits — very large scale integration (VLSI).2
VLSI revolutionized the process of creating complex computer chips by making it more manageable, scalable, and accessible to a broader range of engineers and researchers. The key insight was shifting the focus away from individual components and towards methodologies and tools that enabled designing complete integrated systems on silicon.
While academic systems science struggled (as it still does) to find solid theoretical foundations and unification among its practitioners, PARC was doing applied systems science by using basic systems thinking concepts to solve very practical problems. Their researchers saw how a scattered array of hardware and software inventions could be treated as interoperable coherent wholes and be integrated in order to serve human needs.
Alan Kay, who worked on early portable personal computers, recounts a “blood oath” taken by original PARC founders: “never do a system that wasn't engineered for 100 users."
A Failure to Innovate
When Pake took on his role directing research at PARC, he told his supervisors that five years was too soon to expect research results, but there should be some useful commercial output before ten years.
Pake was remarkably accurate — profits from the laser printer gave Xerox a healthy return on its initial investment in PARC’s first decade of operations. Unfortunately this sort of success was a notable exception, not the rule. Xerox is well known for its failure to appreciate the potential market value of the various other technologies invented by researchers at PARC and commercialize them.
For example, despite support from Pake, Xerox rejected calls to open a personal computer division. Researchers who worked on the early versions of the Alto personal computer wanted to push an Alto III personal computer for mass market consumption, but the proposal was rejected.
Steve Jobs, who had received a $1 million investment in Apple from Xerox pre-IPO and tours of PARC for Apple executives as part of the deal, went on to use the ideas introduced by the Alto in the design of the original Macintosh computer. The Mac successfully, and very profitably, brought the personal computing desktop paradigm to the masses in 1984.
Xerox’s failure to innovate can be attributed to a couple factors worth briefly introducing.
Organizational Structure
In stark contrast with Bell Labs, Xerox was not structured in a manner that facilitated seamless flow from research inventions borne from basic scientific reason to the entrepreneurial innovation required to commercialize and bring products to market.3 4
Bell had specialized research development teams tasked with staying in touch with the cutting edge work being done by scientists and feeding promising commercial technology proposals to systems engineering teams. PARC’s researchers worked in isolation in their ivory towers for five years before a development organization was created in 1976 to take research prototypes and turn them into products.
PARC, located in Palo Alto, CA housed free wheeling scientists creating and working in the office of the future. Xerox corporate headquarters located in Rochester, New York housed corporate businessmen laser focused on ensuring that Xerox’s primary money machine, photocopying, stayed running.
There were high levels of bureaucracy, costly delays in decision making and an overall structure which prevented the easy flow of knowledge across PARC’s borders into the rest of the corporation.
Politics
There were toxic political dynamics at play as PARC researchers feared that the people making critical business decisions knew nothing about technology. Meanwhile many of the executives saw PARC researchers as strange bearded nerds who never left the lab and knew nothing about business.
There was also animosity and distrust between PARC research scientists and Xerox engineers. For example, when there was an attempt to extend knowledge about the GUIs developed by scientists outward to the rest of the company, many of the engineers found the work unintelligible and they couldn’t turn it into a marketable product. Scientists dismissed engineers as “toner heads” obsessed with copying machines while engineers found scientists to be unrealistic and arrogant.
One organizational theorist and civil servant, Myron Tribus, who joined Xerox after serving in the Nixon administration, recounted:
“I was used to the politics at Washington, but at Xerox it was way worse. In Washington, you knew your adversaries and accepted they would work against you. At Xerox, you only found out who was not on your side after you noticed the knife in your back.”
PARC has made for a fascinating case study. While the lab succeeded in inventing the office of the future, its parent company largely failed to innovate and commercialize the technological advances they were privy to.
Instead organizations like Apple, IBM, and Microsoft reaped the benefits and glory of bringing the computing revolution to the masses.
The contrast with Bell Labs who had the basic research to commercialization pipeline down to a science is remarkable. Another useful point of contrast is to note how with PARC, the scientific breakthroughs emerged in the fuzzier realms of computer science, information science, and consumer psychology, not just in the hard sciences as was primarily the case at Bell.
Next week I’ll be exploring Protocol Labs, a name which is probably unfamiliar to readers who aren't engaged with crypto, but one that should certainly command the attention of all systems researchers interested in the intersection of basic and applied science.
Pake, G. E. (1985). Research at xerox PARC: A founder’s assessment: The first director of the Xerox Palo Alto Research Center tells how his management philosophy worked in launching the center and making it a success. IEEE Spectrum, 22(10), 54–61. IEEE Spectrum. https://doi.org/10.1109/MSPEC.1985.6370843
Perry, T. S., & Wallich, P. (1985). Inside the PARC: The `information architects’. IEEE Spectrum, 22(10), 62–76. IEEE Spectrum. https://doi.org/10.1109/MSPEC.1985.6370844
Heracleous, L., Papachroni, A., Andriopoulos, C., & Gotsi, M. (2017). Structural ambidexterity and competency traps: Insights from Xerox PARC. Technological Forecasting and Social Change, 117, 327–338. https://doi.org/10.1016/j.techfore.2016.11.014
On Xerox PARC, and the failure of execution—Only Dead Fish. (2024). https://onlydeadfish.co.uk/2024/04/22/on-xerox-parc-and-the-failure-of-execution/