Palm is one of the most interesting computer companies around at the moment, and that is not just because of its user-friendly products.
The company is a test case of the way in which the computer industry works. Whether it thrives or dies, it will continue to exist in thousands of business studies about what, or what not, to do.
When Palm launched the Palm Pilot, it took exactly the approach that would be expected. As a new company creating a new market, it developed its own hardware, operating system and applications, wrote the manual and shipped products to dealers. It wanted buyers to be able to take the device out of the box and find it useful.
In other words, it took responsibility for delivering the whole product (see Why VHS was better than Betamax).
The Palm Pilot was, of course, a tremendous success. It made Palm the latest example of a vertically-integrated company like, for example, Wang (word processors), Tandem (fault-tolerant computers), Apollo (graphics workstations), Apple (personal computers), or Psion (personal organisers).
The problem it faced was that hardly any vertically-integrated companies survive or keep their independence once a mass-market, horizontal, platform arrives. Palm knew that, in the long run, competition would come not from a single supplier, such as Psion, but from a range of companies sharing a common platform.
All these companies would compete with one another, driving prices down while creating a bigger market. None of them would get 100% of a proprietary pie, like Palm. However, they would, they hoped, get their rightful share of a much bigger communal pie.
Palm was right to be worried, because Microsoft loomed on the horizon. Microsoft learned all about how platforms work from the IBM PC and, since then, has put every effort into trying to repeat its success.
As David Nagel, a former Apple executive who now runs PalmSource, told CNet last year: "Microsoft is the textbook example of how to run a platform business."
It is not clear how much of this strategy Palm articulated at the time. However, Donna Dubinsky had also worked for Apple, and Apple alumni often have one idea in common: "We're not going to make Apple's mistake".
Apple had been the market leader in personal computing before the launch of the IBM PC in 1981, and it launched the Macintosh in 1984. This graphical machine was, in many respects, superior to the IBM PC running Microsoft's character-based MS-DOS, but DOS still wiped the floor with the Mac.
The original Mac had been designed as an appliance, like a Maytag washing machine. As Apple's co-founder, Steve Jobs, said, there were no Maytag user groups. As an appliance, the Mac was a proprietary, sealed box. By contrast, IBM had opted to go with what, at the time, was considered to be an open system, more like the successful Apple II.
The late Philip "Don" Estridge, president of IBM's Entry Systems Division, talked about the background to the PC in a special issue of Byte magazine in November 1983.
By that time, it was clear that the PC had become the market standard. But when developing it, Mr Estridge explained, "we talked about IBM being in a special position to set standards, but decided we didn't want to do anything to introduce standards.
"We tried to do everything we could to understand the existing infrastructure across the board: in marketing, distribution techniques, pricing, customer alternatives, software suppliers, hardware add-on suppliers and peripheral manufacturers.
"We reached that conclusion because we thought that personal computer usage would go far beyond the bounds anyone could see back in 1980. Our judgement was that no single software supplier or single hardware add-on manufacturer could provide the totality of function that customers would want."
Mr Estridge clearly understood the whole product concept, and what it took to establish a platform. For example, he rejected the use of new technology 3.5in drives to replace standard 5.25in floppies because, he said, "you don't need to take on the extra burden of a disruptive medium, no matter how good it is."
That 724-page, ad-packed, issue of Byte, full of products supporting and enhancing the IBM PC, was a testament to the fact that he was right.
The Macintosh was launched a few months later with a stunning "Big Brother" advertisement, which was shown during the Superbowl football game. The ad, an attack on IBM, was wonderful, but sales were so far below projected figures that Apple had to close factories and lay off staff.
According to Michael Malone's study, entitled Infinite Loop, MacWrite author Randy Wigginton collapsed, and spent most of a year in bed watching television. "Everyone who worked there identified totally with their work: we all believed we were on a mission from God," he said. When people didn't buy it, we were majorly depressed."
Thousands of people told Apple exactly what to do to: it had to open up the Mac, license the operating system to rivals, and turn it into a platform that could compete with the IBM PC.
In June 1985, the boss of one small software house, which was trying to become the biggest supplier of Mac software, sent Apple's bosses a memo spelling it out. He even offered to introduce Apple to companies which could be interested in producing Mac-compatibles to compete with IBM-compatibles. But the memo, from Bill Gates, was ignored.
Apple knew that, if it let go of its monopoly, it would not be able to get such high profit margins on its hardware. Anyway, why should other companies be allowed to profit from its ingenuity? It was a catastrophic mistake.
After that, vertically-integrated companies with potentially winning products knew that they had a challenge. Just how did you turn from a vertical structure to a horizontal one enabling other companies to maximise the value of your invention? How did you establish a de facto standard platform when you did not have IBM's huge size and market power?
One obvious answer was to sell out to someone who did have the power, such as IBM or Microsoft, and hope that your ideas did not disappear. Another was to offer to license everything to anyone, which is what Sun Microsystems did with its Sparc processor and SunOS operating system.
However, the most attractive idea was usually to split up the company and bring your supporters on board. If they win, you win and everybody wins.
It is a technique used by two small British companies to reach a global market with what, otherwise, might have remained local products. The first was Acorn, which developed the Acorn Risc Machine (ARM) microprocessor, used in the Acorn Archimedes personal computer. In 1990, the design was hived off into ARM (Advanced Risc Machines), a company co-owned by Acorn, Apple and ARM chip manufacturer VLSI Technology Inc.
Apple used the ARM chip in Newton, its PDA, and now dominates the market for handheld computers and mobile phones. Intel acquired the StrongARM version from Digital Equipment and now has its own implementation, called XScale.
Another British company that said it would not make Apple's mistake was Psion, which pioneered electronic organisers in the 80s. It spun off its Epoc operating system into a new company, Symbian, with the backing of most of the world's giant phone companies. Symbian has not yet enjoyed ARM's success, but it could.
Palm Inc is taking the same approach. It invited competition from cloners, signing up Handspring (a new company formed by Palm's founders), Sony and others to make Palm-compatibles. Palm also tried to avoid trampling on the software houses writing programs for Palm OS, and says it is doing everything it can to grow "the Palm economy".
Finally, it has offloaded control of the operating system to a separate company, PalmSource, which is headed by Mr Nagel.
When I went to see him debate the relative merits of Palm OS against representatives from Microsoft and Symbian at Comdex last November, he grinned wryly and summed up: "Well, we're all selling platforms."
All three are trying to replicate the success of the IBM PC, and the Wintel (Microsoft Windows/Intel Pentium) platform that came out of it. That is still seen as the route to success: until, that is, someone finds a better way.
When interviewed by CNet after his appointment, Mr Nagel said: "It is straightforward how you do an integrated company like Apple. The software and hardware sides are technology and development plays, and the two of them work together to make great products. You can control everything, end to end. You also end up with 2.5% market share.
"I think the great concern, and certainly I shared it, having been in that play before, is that you can't make a better enough product to deal with the superior economics of a more open model. You just can't do it."
It is now up to Mr Nagel to change the mantra from "avoiding Apple's mistake" to "emulating Palm's success".
