Previous | Table of Contents | Next |
Research is about taking risks. Research is permission to fail, as long as you learn and can report on the reasons for failure. A corporation that funds major research efforts has two responsibilities. The first is to monitor the work of its researchers and to understand how that work can be leveraged. The second is to monitor the work of others and to understand how that work can be leveraged.
1979 was a year of ominous events. It was the year Kay did not return. He had gone on sabbatical, visiting Information Sciences Institute, USC. He left Xerox to join Atari. Eventually he went to Apple and then in 1997 joined Disney.
Steve Jobs, then head of the fledgling Apple Corporation, visited PARC. He actually visited PARC twice. The first visit was hosted as a courtesy to the Xerox Venture Group, which had a significant investment in Apple. The visit involved a management demo, orchestrated in the foyer of the entrance to PARC and presented to Jobs, John Couch, and Jeff Raskin (who had visited PARC earlier when he was a UC professor and had encouraged Jobs to see for himself). Ingalls and Tesler gave the demos.8 Then Jobs came back for a second demo. He came back with the entire Lisa programming team. (Lisa was the predecessor office workstation to the Macintosh and the competitor to the Xerox Star Workstation of the same time period.) Jobs sat with the Apple programmers in a conference room and waited. He waited three hours while I, as current LRG manager, argued in vain that a demo for the programming team was an inappropriate give-away of Xerox research results. Worse, Jobs insisted that I give the demo, knowing I was the groups teacher. My manager at the time was head of the PARC Science Center (PARC was split into two while the Corporation sought a new PARC manager who could do a better job of handling the disagreements between CSL and most of the rest of PARC). My manager ordered me to give the demo, claiming later that he was ordered to do so from corporate. It is doubtful that corporate understood enough to be giving any such orders. The plans for Lisa were revised.
8Tesler left PARC that year to join Apple, followed by Ingalls a year later.
By 1979 standard microprocessors were sufficiently powerful to run a language system such as Smalltalk. The NoteTaker ran Smalltalk on Intel 8086 processors. At Stanford, the Sun workstation, part of the larger Stanford University Network project, had been designed using the Motorola 68000 family of microprocessors. The Sun workstation appeared to be the perfect vehicle for Smalltalk. As Peter Deutsch noted, the 68000 was the first commercial microprocessor with a powerful enough architecture (specifically a flat address space of more than 16 bits) to run Smalltalk.
When PARC first started, there was no capital equipment. MAXC (Multi-Access Xerox Computer, a microcoded emulation of a DEC PDP-10) had been built because the request to buy a PDP-10and thereby be compatible with the ARPA community from which most of the research staff camehad been denied. The Altos were built and spread throughout PARC and the rest of the Xerox research community. Then the Dolphins (Xerox 1100) and the Dorados were built. Each time a new machine appeared, battle lines were drawn on whether funding would be provided to recapitalize researchers who appeared already rich in computing power. LRG managed to stay ahead of the funding game by selling old machines to buy new ones (first Altos for Dolphins, then Dolphins for Dorados). Sales were restricted to other Xerox groups, although both the U.S. White House and the CIA (via Xerox Special Interest Systems) were Dolphin users.
Now we wanted Sun boards. We started by borrowing from Fairchild Schlumberger. We borrowed a Sun board and one of Schlumbergers best developers, Alan Schiffman. He and Peter Deutsch set out to design PS (as in Peters Smalltalk, Peter and Schiffman, or Portable Smalltalk). The next step was to sell Dorados to buy workstations from the new Sun Microsystems Corporation. My new manager, the new head of PARC, told me later that CSL was shocked that we were permitted to buy non-Xerox machines. But he supported my argument that someone at PARC needed to understand what the rest of the world was doing.
The PS work needs special mention here because it demonstrated why Smalltalk is so interesting. Smalltalk is a system in which to describe systems. The Smalltalk compiler and development tools were written in Smalltalk itself, lending the system to changeability at all levels. Deutsch designed a 68000 assembler language development environment by writing a 68000 emulator in Smalltalk and using the standard Smalltalk browsers and inspectors to access 68000 code (Deutsch, 1983). Resulting 68000 code was then cross-assembled from the Dorado to the 68000 Sun board. This 68000 development environment was the only example we knew of where one used a combination of subclassing and class parameterization to achieve reusability of an entire development environment as a framework whose concrete completions were themselves environments for different languages. More than half the PS project effort was spent building this highly productive development environment. A similar approach was used to develop 8086 code for the NoteTaker.
PS convinced us that Smalltalk would run well on standard microprocessors. We no longer needed to rely on microcoding Xerox proprietary machines, so we decided it was time to expand the audience for Smalltalk. We decided to create a Smalltalk that the rest of the world could use.
In 1979 we asked Xerox for the right to publish Smalltalk, the language and its implementation and the applications we had built to test the Smalltalk model of computing. Xerox officially gave this permission, remarking that no one inside Xerox wanted Smalltalk. With that simple statement, Xerox officially ignored the increasing demands from a key Xerox customer, the CIA, for Smalltalk and Smalltalk-capable workstations.
Previous | Table of Contents | Next |