Wednesday, September 23, 2009

PARC's Lunatic Fringe

It is well known that Xerox Palo Alto Research Center (PARC), assumed a legendary role in bringing about many of the technologies that we rely on for communication and enjoyment in our daily lives. Most of its early accomplishments in the realm of Human-Computer Interaction were either conceived by or contributed to by the visionary Alan Kay, who enjoys nearly as much renown and mythic stature as PARC itself. Much less known are the side characters whose skill, dedication, and sheer brain power enabled many of these technologies to see the light of day. Soon after Kay joined PARC in 1970, he began to assemble a team called the Learning Research Group (LRG), termed 'the lunatic fringe' by Bob Taylor, the ostensible head of PARC and computer science impresario. Kay both led and depended upon this team of like-minded but otherwise diverse individuals to realize many of his ideas, most particularly as envisioned in his fabled Dynabook.

The Setting
The scene unfolds in 1970, in an intellectually vibrant community about forty miles outside of San Francisco, on land leased from Stanford University, and 3000 miles from Xerox corporate headquarters in New York. Dr. George Pake was invited by the Chief Scientist of Xerox, Jack Goldman, to found a second research center in a location of his choosing, with the only explicit charge being to bring new technologies to Xerox beyond what could possibly come out of the, by then, staid original research center. Funding flowed freely from corporate, enabling PARC to entice top talent who were given unparalleled free reign. Fortuitously, much of the government funding that went to the neighboring Stanford Research Institute (SRI) started to diminish at that time; PARC was there to pick up the pieces.

Kay had already conceived of the Dynabook in 1968, as "A Personal Computer For Children Of All Ages." In 1969 he completed his dissertation which described the FLEX machine, an interactive computer of the future that incorporated many of the features that the personal computer would later have: it had a display screen, was compact in size, and used concepts of object-oriented programming. When contacted by Bob Taylor and informed that several of their former Berkeley Computer Corporation colleagues were regrouping as Xerox PARC, he realized this was his opportunity to see his Dynabook built. For all his desire to share and talk aloud ideas, he tended towards working alone or as head of a small group. He was thus kept apart from some of the other leading engineers, and in due time was head of the Learning Research Group. By 1973, there were eight of them.

The People
Not particularly impressed by academic credentials (one of his own advisors, Ivan Sutherland, once declared "A doctoral thesis is anything you can get three faculty members to sign"), Kay's main criteria for membership to his exclusive group was shared vision. There was Dan Ingalls, who had studied physics and electrical engineering both Harvard and Stanford, but aborted his education before receiving his Ph.D. Kay recognized in Ingalls the intellectual strength and technical abilities he would need to see his projects through. Ingalls brought along Ted Kaehler who had also been working on a less interesting project on speech recognition at PARC. Adele Goldberg was an educational technology specialist who had done work previously at the universities of Michigan and Chicago. Larry Tesler, a graduate of Stanford and previous member of the Stanford Artificial Intelligence Laboratory, joined early, but his time was initially split with another group. Most unconventional was Kay's recruitment of Diana Merry, a PARC secretary who demonstrated strong programming aptitude as evidenced by her abilities with some of the office machines. Each of these individuals played a crucial role in the ensuing years working together under the LRG and branching out beyond.

As the Learning Research Group, they were the advocates for human-centered design at PARC. Though they were focused on education and learnability by children, the ultimate aim was to improve the human condition. They believed that by learning how to program at an early age, children would improve their minds and thereby become more productive and thoughtful adults. While other researchers at PARC, notably the Computer Science Lab (CSL), were preoccupied with technical intricacies and constraints, the LRG was busy envisioning what it would all look like and how it would be received by people. "For CSL the issue was how rapidly they could move data through the machine, whether it was Bravo text or Thacker's design schematics. For LRG it was how to display the same data in the most mind-blowing, dynamic way."

Just as the LRG were the loudest advocates for human factors at PARC, they appeared to be the most fun-loving and cohesive group. Group intimacy was reinforced with retreats to the seaside resort of Pajaro Dunes a couple of hours away where they would spend three or four days not only incessantly talking about their work, but also having musical jam sessions, drinking beer, going on bike rides, and playing tennis. In essence, having adult playtime. In contrast with the other PARC teams, "the lab's ambiance was less like an Army barracks than a bohemian party where all the guests all happened to concur in their host's choice of wine."

The Influences
In understanding learning, Kay largely supported Seymour Papert's constructionism theories. Papert was in turn inspired by Jean Piaget's constructivism. Piaget believed that individuals learned by constructing mental models based on their experiences. In contrast with his contemporaries, he asserted that playing serves a real function for children--that it was central to learning. Constructionist Learning extended these theories, and advanced the idea that one learns best while creating tangible objects in the real world. Papert's LOGO was a simple programming language designed to teach children how to use computer. Instead of dealing with abstract computer language instructions with equally abstract results, the child subjects were able to use a simple programing language to direct a turtle around the screen. Watching the engagement and joy of these children in interacting with the simple computer must have left an indelible mark on Kay's mind, as he has not to this day ceased thinking about children and computing.

Xerox PARC is often credited with the first implementation of the graphic user interface (GUI) as we recognize it today. In fact, with their vision and know-how, they relied upon many existing technologies and created a few new, though key, ones. While at SRI, Doug Englebart developed a system in the early 60's that utilized a mouse, multiple windows, and hypertext. The last, of course, was conceptualized in Vannevar Bush's memex system in 1945. Sutherland demonstrated with his Sketchpad in 1963 one of the first graphic display systems that could be interacted with in real time. With Sketchpad, graphic images could be directly manipulated and in astonishingly complex ways.

The Inventions
One day in a hallway discussion with Ingalls and Kaehler, Kay declared that programming languages did not have to be large to be powerful and committed himself to defining "the most powerful language in the world in one page of code." Within about eight days, he had built the bones for it, based on the ground-breaking principles of object-oriented programming, and the revolutionary Smalltalk was both birthed and named.

Smalltalk needed hardware to operate on. Kay was one of the primary conceptualizers of the Alto, but a PARC engineer from the CSL named Charles Thacker was indispensable in its creation and speed of delivery in just a few months. The Alto was small enough to fit under a desk and designed to serve the needs of a single individual. It was the world's first "personal computer," as coined by Kay. In contrast with existing computers of the time, feedback from any inputs to the machine took immediate effect, so that work with it was refreshingly fast. Kay immediately ordered prototypes of the in-demand Alto for the LRG upon its release. It served as the platform on which much of their research and studies were conducted.

Once the Alto was unleashed, development of Smalltalk charged ahead. Kay did not work in isolation. Initially, he conceived and Ingalls implemented. Later, Merry, Kaehler, and Tesler joined the effort in implementing Smalltalk tools on the Alto. The BitBLT, or "bit block transfer" solution that allowed windows to overlap each other at a processing speed that would be tolerable to users was created jointly by Ingalls and Merry. This was in some ways a missing piece to Englebart's paper paradigm which had inspired Alto's interface and its 8.5 x 11" display and metaphorical desktop. With Smalltalk and BitBLT, windows would be able to be moved over and below each other, resembling the shuffling motion of paper on an actual desk. Pop-up menus, with no comparable equivalent in reality, were invented at this time as well.

In investigating the needs of at least one segment of end users, Kay and Goldberg both brought schoolchildren into the lab and delivered Altos to local schools to test user responses. Smalltalk as installed on the Altos was well-received, and children were soon seen programming with the language within days and weeks. One student devised a paint program over a weekend, having never programmed before. Later, they came to realize that the speed of learning was as much a function of the students' giftedness (and privileged backgrounds) as it was the simplicity of the language. The incredible results were not easily replicated. Still, with the assistance of test subjects in the local gifted program at Palo Alto's Jordan Road Middle School, Kay and his group were able to push the Alto to its bitmapped limits.

The End
It is unclear when the Learning Research Group officially disbanded. Kay and his cohorts on one hand were helping to create a laptop pre-cursor called the "Notetaker"--another rendition of and substitute for the Dynabook (a.k.a. KiddieComp or MiniCom). At the same time, others at PARC were furiously building the Dorado--which was to be the most powerful computer of its time, regardless of size, and therefore the antithesis of the Notetaker. A few years ahead of where they needed to be according to Moore's law, in the end portability was sacrificed for processing power, and the Notetaker was produced as a 45-pound beast.

Meanwhile, Smalltalk was spinning beyond Kay's original conceptions of it. He found himself increasingly dissatisfied, and wanted to rewrite the language from the ground-up, but the others at LRG had become too emotionally invested to abandon their work and start anew. This indicated a loss of shared vision to Kay. He left PARC in 1978, and Goldberg assumed leadership of the LRG. She led Smalltalk through to its next incarnation, Smalltalk-80, the first public release.

According to legend, in 1979, Steve Jobs and an assembled crew of Apple programmers, with the support of Xerox's top management, stormed onto PARC's premises and demanded to be shown PARC's new technologies. The researchers designated to perform the demo were all members of the LRG. Reluctant to "give the kitchen sink away," Goldberg argued for hours with superiors against revealing their trade secrets. Xerox had a stake in Apple at the time, and she was overruled and dispatched to give the presentations. Thus the Apple Lisa was born, incorporating many of features of PARC's home-grown GUI. Jobs was so blown away by the user interface features of Smalltalk on Alto that he failed to notice two other pioneering achievements--object-oriented programming and networking. One of the most important takeaways Apple did glean from the visit, however, was that a computer should be fun to use.

The Legacy
Though Kay would claim that his Dynabook in its pure form has not yet been realized, due to shortcomings in key software and educational curriculum, the work of the LRG in the areas of graphic user interfaces and object-oriented programming has had an undeniably lasting impact. The interface of the Alto inspired the Apple Lisa, which transmuted into the Mac OS, that was then swiped by Microsoft for Windows. In many ways, the LRG gave birth and significance to the idea of the user experience. Smalltalk continues to be used to this day, but more importantly its concepts have inspired a plethora of other languages. Kay has remained true to his principles, with his role in the One Laptop per Child program: still hoping that the illustrations he had created in his relative youth--that depicted children in bucolic settings with their portable computers learning and playing--would come to fruition.


Hiltzik, Michael, Dealers of Lightning: Xerox PARC and the Dawn of the Computer Age (HarperCollins, New York, 1999)


* Constructionism
* Constructivism
* graphical user interface
* history of the graphical interface
* desktop metaphor
* Xerox Alto
* Dynabook
* SmallTalk
* Adele Goldberg
* Alan Kay
* Dan Ingalls
* Diana Merry
* Larry Tesler

~ Jenny Wang

No comments: