|"The Fallen Apple" Corrections
By Larry Tesler
April 11, 1999
Apple Confidential: The Real Story of Apple Computer, Inc. by Owen W. Linzmayer, No Starch Press, 1999.
This book is a good account of the story of Apple, but at least some details are quite wrong.
I'll stick to what I know first hand. The book's account of my involvement with Newton is garbled, and the chronology is wrong in many respects.
I did not beg John Sculley to let me replace Steve Sakoman as the Newton project leader. After Sakoman resigned, I offered to spend a few weeks as acting manager to assess the skills of the team, something Sculley had been planning to do on his own.
I prepared the team for the Board presentation that led to the project's continued funding. After the Board presentation, I helped the group to seek a new manager. When none could be found, Sculley asked me if I would continue as manager. I said yes with mixed feelings.
As VP of ATG, I had been Marc Porat's and Bill Atkinson's manager, and the executive champion of what became General Magic. Both before and after that pivotal Board meeting, Sculley and I tried to get Porat's project and the Newton to combine. But the personality clashes and differences in vision were insurmountable. We spun out General Magic instead.
During the assessment period before the Board meeting, I urged the Newton team to bring the target price down from $6000+ to $4000, then to $2000 or less. They accomplished this partly by tossing out two of three microprocessors, partly by replacing the radio LAN by infrared, partly by compromising on features such as battery life and screen brightness, and partly by reducing the size.
I was at no time an advocate of a "large tablet" costing "$5000". Work on such a beast stopped as soon as I arrived. I wanted as small and cheap a device as possible. But I did not think Apple could manufacture a pocket-sized device, so I compromised on an in-between size and a $1500-$2000 price.
The original three processors were two AT&T Hobbits and an ARM. Apple had paid not one, but several million dollars to AT&T for the Hobbit processor. The Hobbit was rife with bugs, ill-suited for our purposes, and overpriced. We balked after AT&T demanded not one but several million more dollars in development fees. Instead, we invested $2.5M in ARM, less than what we would have paid to AT&T. After ARM's 1998 IPO, Apple's financial return on that investment was something like 1000%. [Note added 10-Oct-2004: It ended up being more like 30,000%. Apple made an $800 million in profit on a $2.5 million investment. That profit exceeded the total cost of the Newton project.] The Hobbit's commercial demise a few years later validated our decision. The processor switch did not "create a lot of extra work", since relatively little processor-specific code existed at that time, and much of that was unsuited to the new system design.
The general story line is correct about Steve Capps and Michael Tchao's unflagging efforts to find a way to make Newton a pocket-sized device. They were actually a trio: the third advocate was Mike Culbert.
We had to secure an exemption from Apple manufacturing to get a Newton designed and manufactured outside. Sculley himself helped us to make a deal with Sharp. Once we had secured the exemption and lined up Sharp, the main difference of opinion between me and Capps/Tchao was whether to run two projects in parallel or just one. Unconvinced that the pocket-sized Newton would be adequate, I refused to cancel the mid-sized version. The project had already changed direction several times. I was concerned that we would never finish anything we started.
Capps broke the impasse by (as he later admitted) intentionally understating, by a large factor, the amount of work it would take to implement the pocket-sized Newton. Based on his assurances, I cancelled the mid-sized Newton. Sculley and I persuaded Mac engineering to cancel Paul Mercer's Pocket Mac, an evolved version of a device that had been developed by Harry Vertelney's group in ATG with my encouragement. It was Capps' intentional underestimate that led to Apple's overly optimistic product announcements and the subsequent rush to ship a product before it was really working.
On the other hand, Capps was right to want to toss out the overdesigned Ralph/Dylan language in favor of the similar but far more practical NewtonScript, a fact that the book does not even mention. One mistake I had made was to ignore one good bit of advice from Sakoman: I let ATG manage Ralph/Dylan, even though Newton was the only customer. The result was a mismatch between the goals of the language designers and Newton's needs.
Another omitted story is that early on, I hired a business anthropologist, Eleanor Wynn, to interview customers and find the right application for Newton. She concluded it would sell best to police and fire departments, medical clinics, and other vertical markets. Tchao, Capps, and others did not want to hear this. They rejected the study, ran their own focus group, and concluded that the Newton should be a Sharp Wizard-like organizer for mobile professionals. As it turned out, Eleanor's study was right on the mark. She had accurately predicted the eventual market in which the Newton was most successful.
There were consistent messages from both studies, though: (1) a cell phone should be incorporated; (2) natural handwriting was essential--the customers were adamant that they would never learn a stylized script. We ignored (1) because it was then uneconomical. We believed (2) and struggled to make handwriting work, but the PalmPilot later proved that customers would indeed learn a stylized script.
A draft of an early Newton flyer said "It recognizes your handwriting". I insisted that the phrase be removed, and the copy writers promised to take it out. The printed flyers did not arrive at the CES show until minutes before we started the presentation. There, on page 1, it said, "It recognizes your handwriting". The PR team must have thought that I would forgive this transgression after I saw what wonderful press coverage we received from such a bold statement. But what they had unknowingly done was to hammer a large nail into the Newton coffin by setting expectations we knew we could not meet.
Also omitted is the story of how Alan Kay tried to convince the Newton Group under Sakoman to make the Newton a wireless accessory to a desktop computer, automatically synchronizing data when it approached its host. This was basically the PalmPilot concept, three years before Palm Computing was founded. When I took over the Newton Group, I invited Kay several times to return and sell the concept to the group. He declined on the grounds their opinions were rigid and they would never listen.
Book authors like heroes and villains. But we all made a lot of good calls and a lot of terrible mistakes during the Newton project.