Accidental Empires, Part 13 -- All IBM Stories are True (Chapter 7)
Thirteenth in a series. If you ever wondered about the real story behind the IBM PC, this chapter of Robert X. Cringely's 1991 classic is the one for you.
I live in California in a house that I can’t really afford in a neighborhood filled with blue-haired widows and with two-earner couples who have already made the jump from BMW to Acura and in their hearts are flirting with voting Republican.
Remember when life came mainly in black and white, and Wally and the Beav walked down a street as the credits rolled across them? That was my house they walked by on that tree-lined street, my 50-by-105 foot lot, my gnawing termites, my 1957 Studebaker Golden Hawk dripping oil in the driveway, and my orange tree dropping oranges in the side yard. For a kid raised in Apple Creek, Ohio, walking out the door in the middle of winter and pulling a fresh orange off your own tree is heaven.
And New York City is hell.
The New York I reluctantly visit is filled with angry people, garbage, burned coffee, potholes, and overcooked vegetables. Yet I have lots of friends who live in Manhattan and tell me it’s the most wonderful place. Their New York has theater, music, museums, a great library, and Central Park, while all I have at my house in Palo Alto is good phone service, one orange tree, and a bookshelf containing the complete works of Louis L’Amour. My friends are certain they would die of boredom in Palo Alto, while I’m just as sure I’d die in New York, though not of boredom.
Who’s right? We both are.
There’s a psychological principle called cognitive dissonance at work here. When the reward (having an orange tree) is out of proportion to the effort required to achieve it (a $3,000 monthly mortgage payment), I am put in a state of dissonance, which can be resolved only by selling the house and moving back to Apple Creek or by warping my entire sense of values to convince myself that it’s worth all the effort.
Of course, I choose to warp reality. We all do. I tell myself that warm winters and fresh oranges are worth anything, while my friends in New York say exactly the same thing about plays that they don’t really attend and parks they are afraid to visit.
Cognitive dissonance plays a major role in all of our lives, and so far it’s the only reason I can come up with that most people continue to work for IBM.
Here’s what I mean. In early 1983, a big guy named Don Estridge was about to become IBM’s vice-president of worldwide manufacturing, making about $250,000 per year in salary and bonuses. What does a vice-president of worldwide manufacturing do, you ask? Not much in a company like IBM, where each division does its own manufacturing while fighting with all the others. Estridge’s VP role was really a holding pattern designed to stash a guy who had served the company well but was now out of sync with the Brooks Brothers reality of what the largest computer company in the world was all about.
Since 1980, Estridge had been head of the team that invented the IBM PC, moving Big Blue in ways and at a speed that it had never been moved before. Estridge had been told to go out and break the rules, did as he was told, and by doing so became a dangerous man. Once his team had succeeded by specifically being as unlike the rest of IBM as possible, Estridge could no longer be trusted because his ways were no longer the ways of IBM. Welcome to the vice-presidency of worldwide manufacturing.
Then one day in early 1983, 28-year-old Steve Jobs showed up on Don Estridge’s doorstep in Boca Raton, Florida, looking for a new president for Apple Computer. Jobs has only three ways of dealing with people: he seduces them, castigates them, or ignores them. In fact, everyone in Jobs’s life eventually runs through all three modes, sometimes more than once. Jobs was in seduction mode when he approached Estridge, and Steve can be incredibly seductive when it suits his needs. He is the best salesman in the world and put all those talents to work trying to recruit the tall man from IBM.
Estridge, who had for years programmed his own Apple II at home, took a trip out to Apple headquarters in California and liked what he saw. The money wasn’t bad either: $1 million annual salary, $1 million signing bonus, and a $2 million loan to buy that Silicon Valley dream house.
Estridge was about to be stashed in a corporate backwater and would probably never be considered for the top job at IBM. The fit with Apple was nearly perfect, and the money was terrific. Nevertheless, he agonized for a while, talked it over with his closest friends, and turned it down: cognitive dissonance.
Don Estridge couldn’t have left IBM even for the top job at Apple. That’s because Apple is only a company, while IBM is a country.
With annual sales around $60 billion, IBM has a greater gross national product than most countries. It has a relatively stable population of around 380,000 workers. Throw in the spouses and their 1.8 kids each, and we’re looking at more than a million citizens of IBM.
Demographically, IBM is most like Kuwait, but temperamentally IBM is Switzerland. Like Switzerland, IBM is conservative, a little dull, slow to change, yet prosperous. Both countries are in the habit of taking in more money than they give out. Both countries learn slowly and adapt at their own pace. Switzerland and IBM can survive anything, or at least think they can. They may be slow, but don’t mess with them, because they will fight to keep what is theirs. And if pushed, they’ll fight dirty.
Like Switzerland, IBM is landlocked, though Big Blue’s barriers are regulation and internal rivalries, not geography. Instead of facing Austria, France, Germany, Italy, and Liechtenstein on its borders, IBM is surrounded by U.S. antitrust laws and a 1956 consent decree that somewhat limits its ability to wreak havoc upon the land. Even more limiting is the rivalry between IBM’s different computer divisions, each protecting its turf from incursions by the others. IBM’s mainframe division worries as much about competition from the top end of the company’s own minicomputer line as it does from any outside competitor. And there is no law or consent decree limiting the amount of infighting that can go on within the company.
The citizens of IBM didn’t invent the computer. They don’t make the most powerful computers either. The citizens of IBM just make more computers than anybody else. So just as Levis define blue jeans to a world that somehow survived Gloria Van-derbilt, IBM defines computers.
IBM computers don’t stand apart, but IBM people do. Of all the companies I’ve dealt with, the only two whose people consistently present a common front, a kind of unique company style, are those from IBM and Procter & Gamble. This comes from their hiring practices, the way they indoctrinate their workers, and the fact that both companies have had official songbooks. There must be something very unifying about getting together with a thousand other folks at a sales meeting in New York or Cincinnati and singing your guts out in praise of the Old Man.
The men and women of IBM have their own language and stick to it. A minicomputer is a mid-range system. A monitor is a display. A hard disk drive -- a data storage device that has one or several magnetic platters and spins continuously at 3,600 rpm -- is for some reason called a fixed disk, although it isn’t fixed at all. Sticking to these terms preserves the illusion that IBM’s $800 display is somehow different from Samsung’s $349 monitor or that IBM’s fixed disk drive, made under contract by Seagate, is somehow superior to the exact same drive sold for half price under the Seagate brand.
Like Rolex or Gucci, the people of IBM know that they are not really selling computers at all but the IBM name.
IBM people are a little smug, a little slow, and slightly overweight. Most IBMers are hired straight out of school and have never worked for another company. They are folks who drive Buick Regals and take them to the car wash every Saturday morning, paying extra to get the hot wax. Their contented middle-class style bugs the hell out of Silicon Valley entrepreneur types, who want to do business with IBM and yet can’t understand that there are folks in the world -- even in the world of computers -- who aren’t, like them, madly driven to have a fortune and a Ferrari before their midlife crisis.
IBMers aren’t in the business to become millionaires. How can they? These people are not sitting on stock options in some start-up, waiting for their penny shares to go public at $8. They work for a company that went public more than sixty years ago -- the quintessential blue chip. Even IBM salespeople, who work on commission, selling computers that can cost millions of dollars, have carefully set quotas that effectively limit their earning potential.
Ross Perot, founder of Electronic Data Systems, was one IBM salesman who got fed up and left the company when he filled his sales quota for the entire year before the end of January and knew that he wouldn’t be allowed to sell any more computers -- or earn any more money -- for eleven more months.
The people of IBM don’t need to be rich. They either want the security of working for a company that will employ them for life, offering fringe benefits beyond those of any welfare state, or they want the sheer power that comes from eventually working up into the stratospheric reaches of the most powerful company on earth. Money and power are not synonymous at IBM, where power is preferred.
The price of both prosperity and power is compliance with the rules and the pace of IBM. The rules say that you go where the company asks you to go, do what the company asks you to do, and don’t talk about work with strangers. There is a class of company that won’t tolerate different behavior, and those companies sometimes suffer for it. IBM is like that. The pace is slow because it takes time to get 760,000 legs marching together.
Every IBM employee’s ambition is apparently to become a manager, and the company helps them out in this area by making management the company’s single biggest business. IBM executives don’t design products and write software; they manage the design and writing of software. They go to meetings. So much effort, in fact, is put into managing all the managers who are managing things that hardly anyone is left over to do the real work. This means that most IBM hardware and nearly all IBM software is written or designed by the lowest level of people in the company -- trainees. Everyone else is too busy going to meetings, managing, or learning to be a manager, so there is little chance to include any of their technical expertise in IBM products.
Go back and read that last paragraph over again, because that’s why IBM products often aren’t very competitive.
IBM has layers and layers of management to check and verify each decision as it is made and amended. The safety net is so big at IBM that it is hard to make a bad decision. In fact, it is hard to make any decision at all, which turns out to be the company’s greatest problem and the source of its ultimate downfall (remember, you read it here first).
Except at the very highest positions in IBM, this corporate support system produces a class of executives with bovine, cud-chewing dispositions, who think only on command and typically rely on the company to tell them what to do and when to do it. Before beginning each new assignment, for example, IBM people are thoroughly briefed with all the information the company believes they will need to know in order to do their new job. The briefings are so complete that most IBM people don’t bother to do any outside reading or research of their own. If IBM marketing executives know how their personal computers compare with the competition’s, it is nearly always through their briefing books and hardly ever by actually using the other guy’s hardware -- or even their own.
And at the top of IBM, where synapses do pop on occasion, and brain activity is usually, though not always, measurable, nearly all of that activity goes into playing corporate political games, as though competitors and even the global computer market didn’t exist.
It was corporate infighting, in fact, that finally made entry into the microcomputer market so attractive to IBMers grown tired of slugging out the next point of mainframe market share while at the same time engaging in internecine warfare with other company divisions. In the microcomputer business, there looked to be no divisional rivalries to worry about, no antitrust considerations, and, most imponant, the customers were all new, fresh meat, having never before felt the firm handshake of an IBM salesrep. Every sales dollar brought in to buy a microcomputer would be a dollar that would not otherwise have come to IBM. There was something pure about that, and the IBM executives who led the company’s assault on the microcomputer market knew that success on this new battlefield could eventually lead them to the real font of power: IBM worldwide headquarters in Armonk, New York.
There were lots of players in the microcomputer business back in 1980, but Apple, Atari, Commodore, and Radio Shack all looked about the same to IBM -- small. Total U.S. microcomputer sales had reached $1 billion, so there was a market worth dominating. And as Apple had in 1977, IBM saw the market being sales to small businesses, a segment that the company had previously touched only through its typewriter operation.
The IBM Personal Computer that eventually came to market in late 1981 came from a renegade independent business unit based in Boca Raton, Florida. This wasn’t IBM’s first try at developing a microcomputer. At least four other designs had been proposed to management in Armonk, including one earlier design from Boca. The major difference between the project that eventually produced the IBM PC and these earlier efforts was that the group of men brought together in July 1980 by Entry Systems Division (ESD) lab director Bill Lowe were pledged to do their work in real time, not IBM time. They had just one year to bring their product to market.
A year is no time at all to IBM. At the time that Lowe got the go-ahead for Project Chess, which would produce Acorn, code name for the IBM PC, there was an ESD project called Datamaster entering its fourth year of development with no end in sight. Datamaster was an attempt to add some microcomputer functions to IBM’s Displaywriter dedicated word-processing system. If it took IBM four years just to throw additional functions into an existing product, building an entire computer system in one year must have looked impossible.
It was impossible, and Lowe knew it. There was no way that IBM could develop a personal computer in a year. The best they could do was to gather hardware and software from other companies, get them to work together as a system, and then slap an IBM label on the outside. That’s what Bill Lowe decided to do.
The question was whether to build or buy. Up to this point, whenever IBM had been faced with the choice between building a component in an IBM plant or buying it from an outside supplier, the decision was always build, build, build. This was because power within Big Blue was measured in part by the number of workers under each manager. Workers at some supplier’s plant in Bayonne didn’t count.
But Lowe and his crew, breaking their first of many rules, decided to buy everything. They started by looking for software. Since Lowe wanted to buy his operating system software from an established vendor, CP/M looked like his only choice. CP/M came from Gary Kildall’s Digital Research, only for some reason IBM didn’t know that. The usually infallible briefing book said that CP/M was a Microsoft product. In probably his last gracious gesture toward a competitor, Bill Gates told the caller from IBM that a mistake had been made and gave them Kildall’s number in Pacific Grove.
There was still room for IBM and Microsoft to do business, though, because the other software component that seemed to be required in a 1980 microcomputer was a built-in version of the BASIC programming language. Apple, Commodore, and Radio Shack all came with built-in BASIC so users could write their own simple programs. If Acorn was to compete successfully against these machines, IBM would need a BASIC, too. Microsoft was the oldest and best-known provider of microcomputer BASIC and was IBM’s probable vendor choice. It didn’t hurt, either, that Mary Gates sat on the national board of United Way along with IBM chairman John Opel and that the two had become friends. Opel was impressed that Lowe was talking with Microsoft and said so, cinching the deal.
Proposing to build a new line of computers around the products of a couple of five-year-old software companies from the West Coast was a bold move for Lowe and a risky one. Jump, for a moment, into Bill Lowe’s shoes. It’s July 1980. The all-powerful IBM Corporate Management Committee (CMC) has just heard his bold proposal to enter the personal computer business within a year. They tell him to come back in a month with details. The whole plan depends on getting reliable suppliers, so Lowe sends his lieutenants out to Digital Research and Microsoft to find out what kind of people these are. When the IBMers arrived in Pacific Grove, California, to talk with Gary Kildall at Digital Research, he wasn’t there. Despite his appointment with IBM, Gary had gone flying in his small plane. Not a good first impression.
With Gary out flying around, the people left in charge at Digital Research didn’t know what these IBM guys wanted to talk about, and the IBM guys wouldn’t talk about anything until a nondisclosure agreement was signed.
Remember that IBM operates at all times under the consent decree of 1956, which limits its ability to compete. Remember, too, that IBM has the largest legal staff of any U.S. corporation, devoted primarily to finding ways to turn what ought to be limitations into advantages. Enter the IBM nondisclosure agreement, which is the legal equivalent of a neutron bomb, destroying only the people but leaving their technology intact.
Nondisclosure agreements place limits on the ability of parties to reveal the secrets of organizations with which they are doing business. IBM’s standard nondisclosure agreement goes even further. By signing the IBM agreement, would-be suppliers agree that whatever they tell IBM is not confidential, while whatever IBM tells them is confidential. In other words, while the IBM guy on the other side of the table can tell anyone at all what you reveal to him about your company and its plans, he can take you to court if you repeat one of his jokes, much less reveal any IBM secrets. And if IBM takes legal action, the agreement prohibits the other party from even offering a defense.
IBM’s Big Brotherish explanation of its nondisclosure agreement is that would-be suppliers can protect their secrets simply by not revealing them to IBM representatives. In fact, the agreement urges signers not to reveal anything that is confidential. But how can companies do business without revealing confidential information? They can’t. Little companies that want to do business with IBM must sign the we-win, you-lose nondisclosure agreement, bare their corporate souls, and pray that an eventual IBM contract makes it all worthwhile. Big companies that talk with IBM throw their own nondisclosure agreement on the table and demand that IBM sign it too.
Jump back to Pacific Grove, where Digital Research didn’t even have a nondisclosure agreement of its own. Gary was still flying around somewhere over the Santa Cruz mountains, while Dorothy Kildall squinted at the IBM nondisclosure agreement, imagining her new house with its stable and hot tub going on the auction block following an IBM legal action. She refused to sign, so the men from IBM left town, having never revealed their plans for Acorn but still needing an operating system.
The IBMers headed north for their meeting at Microsoft headquarters in Bellevue, Washington. At this time, Microsoft had about fifty employees and was selling versions of the FORTRAN and COBOL computer languages in addition to its many varieties of BASIC. Microsoft had a hardware division, too, that produced a circuit card that made it possible for Apple II computers to run the CP/M operating system, which at that time had better word processing and database programs than were available on the Apple. Microsoft’s Softcard, and the operating system software that shipped with it, made Bill Gates at that time the largest seller of Gary Kildall’s CP/M. Gates knew what a good business selling operating systems could be.
Gates, Paul Allen, and Gates’s buddy from Harvard, Steve Ballmer, put on neckties for a change and met with the IBM crowd. They shrugged and signed the nondisclosure agreement, enjoyed a short period of bewildering small talk, gave a tour of the building, and saw the IBM contingent to the door. Nobody from IBM mentioned that the company was working on a personal computer; that came in the next meeting a few days later.
At their second meeting, IBM asked Microsoft to supply a BASIC language for its new computer. According to Gates, the design they described to him was for an 8-bit computer similar to any of the many CP/M systems already on the market. Gates urged the IBM group to go instead with one of the new 16-bit processors just being released. Using a 16-bit processor would make the IBM PC seem more powerful than its competitors, and it would allow the machine to use more memory too. Apple IIs at that time were limited to around 48,000 characters, or “bytes” of memory—usually called 48K. Most CP/M machines hit the wall at 64K. A 16-bit processor could address vastly more memory than competing machines, offering a marketing advantage and a clear growth path for the future.
The argument was persuasive, but IBM still didn’t have an 8-bit operating system, much less one that would run on 16-bit machines. Gates said he might be able to help out with an operating system.
Choosing a 16-bit processor was easy. Intel, Motorola, and National Semiconductor were all shipping 16-bit processors at the time. Intel had the 8086 and 8088 processors. Motorola had the 68000, and National had its 16032. The National processor was elegant and powerful; the Motorola was powerful and easy to write software for; the Intel 8086 was fairly powerful but had an awkward memory architecture; the Intel 8088 was an 8086 without the power.
Of course, IBM chose the 8088 -- the least attractive of all the processors from a technical standpoint. In this case, technical considerations took a back seat to IBM’s manufacturing and marketing concerns. The plan was to build a computer without any custom components -- just off-the-shelf parts from major semiconductor makers. The 8088 was the only 16-bit processor for which there was available a full complement of the support chips required to build a computer. Motorola and National were still working on their 16-bit support chips, as was Intel for the 8086. But the 8088 was a 16-bit processor in an 8-bit body, since it used an 8-bit data bus -- sending and receiving data 8 bits at a time and then processing them in 16-bit mode. This 8-bit bus is what made the 8088 less powerful than the other contenders, but it also made it possible for the 8088 to use support chips intended for the earlier 8080 family of Intel 8-bit processors. Since the 8088 was the only processor that could be used without developing custom support chips, it was the only processor that fit IBM’s needs.
The 8088 was cheaper than the other processors because it could use the older support chips and because Intel had deliberately priced it below the 8086. But price was not a major consideration in IBM’s decision.
The fact that the 8088 wasn’t as powerful as the other processors was actually seen by IBM as an advantage, since it meant that Acorn would not draw customer attention from the company’s mid-range systems, which were only slightly faster in many applications, though they cost thousands more. Although Project Chess was going ahead without input from IBM’s other divisions, Lowe knew better than to ask for trouble, so the embryonic PC was made deliberately slower than it might have been. It would have been fairly easy too to give Acorn the ability to emulate any or all of IBM’s terminals so the new PC could be used to communicate with IBM minicomputers and mainframes. But that sales advantage was deliberately avoided because it would have killed the company’s terminal business.
But wait. What about the operating system?
At the moment IBM was having its second meeting with Bill Gates, there were no 16-bit microcomputer operating systems on the market. If IBM had waited for Gary Kildall to get back from the airport, it might have learned that Digital Research was already working on CP/M-86, which would run on Intel’s 16-bit 8086 microprocessor and on its little brother, the 8088. CP/ M-86 would be ready to go about the time that IBM planned to release its personal computer, too, so it would have been a logical choice, had IBM known that CP/M-86 existed. As the largest seller of CP/M software, Microsoft knew that Digital Research had CP/M-86 coming down the chute, yet Gates, Allen, and Ballmer never mentioned it in their second meeting with IBM. Remember that the IBM nondisclosure agreement specifically urged them not to reveal any confidential information. CP/M-86 was clearly confidential.
Gary Kildall thought that he and Gates had divided the software market, with Digital Research taking the operating system business and Microsoft controlling the programming languages. Bill Gates knew better.
Across Lake Washington, at a company called Seattle Computer Products, was the operating system that Bill Gates wanted to sell to IBM. All he had to do was get it.
In business, as in comedy, timing is everything. There was nothing about QDOS, Seattle Computer Products’ 16-bit operating system, that couldn’t have been created just as well by programmers at Microsoft. But Microsoft programmers hadn’t created it, and Tim Paterson of Seattle Computer Products had. QDOS, which stood for “quick and dirty operating system,” was a 16-bit clone of CP/M intended for an 8086-based computer being developed by the small company. All QDOS commands were the same as in CP/M. Paterson admitted to a little “low-level borrowing” from CP/M, too, but claimed that most of the code was his own.
Gary Kildall still thinks a lot of the QDOS code was stolen straight from his CP/M. “Ask Bill why function code 6 [in QDOS and still in MS-DOS, more than ten years later] ends in a dollar sign. No one in the world knows that but me”.
There was nothing earthshaking about QDOS, except that it already existed. Bill Gates was buying time more than anything else when he paid Seattle Computer Products $50,000 for rights to the operating system. It must have seemed like a lot of money at the time.
Here’s a great scene that never happened. Bill Gates flies to Florida, makes his pitch to IBM, offering to sell it a product called Dirty DOS, that, by the way, has at least some code stolen line for line from CP/M. The ears of Justice Department lawyers 900 miles away would have perked up. The IBM legal department, which was then suing Fujitsu for stealing IBM code, would have had a corporate seizure. And young Bill Gates would have found himself standing in the sun-drenched IBM parking lot wondering if it was something he said.
Instead, when Gates made that flight to Florida, he kept to the letter of IBM’s own nondisclosure agreement and didn’t reveal much about the true heritage of QDOS, now called MS-DOS, other than that it had been developed with the help of Seattle Computer Products.
Not everyone at Microsoft was so certain that the company ought to get into the operating system business. Microsoft was already running at full capacity just doing languages, so adding QDOS would require a major expansion. What if Microsoft expanded and then IBM canceled Project Chess at the last moment? Big Blue had already canceled four other microcomputer projects. If IBM canceled the deal and Microsoft couldn’t find other customers for QDOS, then that $50,000 paid to Seattle Computer Products really would have been a lot of money. So while Bill Gates has a well deserved reputation for being cheap, his caution in acquiring QDOS was not unfounded. By jumping into bed with IBM, Gates was putting his entire company at risk, and he knew it.
In short order, Microsoft and IBM concluded a co-development agreement making Microsoft responsible for all the system software for Acorn. Gates, Allen, and company would finish the development of QDOS, which would be sold by IBM under the name PC-DOS. They would also provide a BASIC interpreter that would be shipped with each machine in read-only memory. Never before had IBM allowed itself to be so dependent on a single supplier, much less one run by a 25-year-old who ought to wash his hair more often. Aligning with Microsoft was a daring move for Lowe’s crew from Boca Raton and a clear indication of how independent they really were from the old way of doing things at IBM. The Microsoft connection made the IBM PC possible, and the IBM connection ensured Microsoft’s long-term success as a software company.
Bill Lowe came up with the idea of IBM’s building a personal computer. He created a renegade design group, mapped out a design for the original IBM PC, found Microsoft to provide the system software, and sold IBM’s Corporate Management Committee on the project. Lowe’s reward, which came in late 1980, was a promotion to vice-president in another division of IBM. Lowe’s bosses at IBM saw building the organization, rather than the PC, as the object of their enterprise. It did not matter, then, that promoting Lowe meant taking him away from his brainchild; the company was built on large teams and the idea that no one person could be critical to an IBM venture. And Lowe, who had worked at IBM for eighteen years, toed the organization line by gratefully accepting the promotion to run IBM’s lab in Rochester, Minnesota.
In most other PC companies, the person in Lowe’s position would have been kept on the project until it was complete. Any change -- even accompanied by a promotion -- would have been viewed as a sign of disfavor. But Bill Lowe was a company man and went where the company said he should go.
Lowe’s hand-picked successor was Don Estridge, who carried out Lowe’s vision for Acorn. Most of this vision was defined not by how the computer would be but how it would not be. For one thing, it would not be sold by only IBM’s direct sales force because there was no way that a company geared to direct sales of million-dollar mainframes could use the same sales force to sell microcomputers profitably to end users at the prevailing $3,000 price point. Most Acorns would be sold in retail stores. The first computer retailers to carry IBM products were the ComputerLand chain and Sears Business Centers -- Sears Roebuck & Co.’s plunge into the office equipment business that would coincide with the IBM PC’s introduction.
IBM was deliberately bringing a higher, and more boring, level of professionalism to the business of selling computers. You can imagine that the first visit of Sears representatives to Boca was very different from their first visit to Atari’s computer plant in Sunnyvale, California. When the Sears buyers came to the Atari plant, they found a typical Silicon Valley tilt-up building filled with noisy production machinery, rock music, and pot smoke. Atari founder Nolan Bushnell gave the Sears crowd a tour by sitting them in cardboard cartons and literally running them down the assembly line.
IBM did things differently.
In order to meet its price point, profit margins, and delivery schedule, a lot of features were left out of the basic PC. The base machine, for example, would ship without serial or parallel ports for adding telephone modems or printers. No video graphics capability was built in. And while QDOS could theoretically address up to 640K of memory, the basic machine came with only 16K, which could be expanded to a maximum of only 64K by switching to higher-capacity memory chips on the main circuit board, called the motherboard. It was possible to add serial and parallel ports, graphics capability, and memory beyond 64K to the PC, but that required adding extra circuit cards that fit in special slots on the motherboard.
The Apple II had used circuit cards to add memory and features because such cards were the norm in HP 3000 minicomputers used by Steve Wozniak in his earlier job designing hand-held calculators at Hewlett-Packard. Woz designed his own data bus, or scheme for adding special function cards to the Apple II. Most CP/M machines used another bus, called the S-100, for adding special features and memory. An Apple II card would not work in an S-100 computer, nor would an S-100 card fit in an Apple II. There were many manufacturers of Apple II and S-100 cards because both bus standards were published by their inventors, with full information available about how to design a card to work in each type of computer. IBM engineers came up with yet another bus design for the PC, and like the other companies, they published their bus specification so that third parties could design cards to go in IBM PCs.
IBM’s rationale for publishing its bus specification was a good one. For one thing, the consent decree of 1956 required that they publish the technical specifications of all their products, though that was one rule that had occasionally been overlooked in the past. More important, with only a year to build Acorn, there just wasn’t enough time to develop many add-in circuit cards, so IBM would do serial and parallel cards and a video card, but most other such cards would have to be left to third parties to develop. Also, IBM was going to have a hard enough time making acceptable profit margins on Acorn itself, and an analysis of the add-on card business looked even worse, so the decision was made to leave those crumbs to outsiders. Other companies would have to assume the development and marketing costs of add-on cards, but the existence of such cards would only help sales of IBM PCs.
Had anyone in Boca bothered to notice, Acorn was not going to be a very proprietary machine. Microsoft retained the rights to sell QDOS to companies other than IBM. Every component in the new machine came straight from some semiconductor company’s stock bins, with the exception of the ROM-BIOS chip, which linked IBM’s hardware to Microsoft’s operating system software. The bus specification was published and available to any manufacturer who wanted to implement it. All of this meant that there was not much keeping other computer companies from building computers exactly like the IBM PC, piggybacking on Big Blue’s probable success in the microcomputer market. All that a would-be clone maker would have to do is reverse engineer the ROM-BIOS, something that Amdahl Corp. was already doing in the IBM mainframe world and that the courts had decided was legal. But IBM was not worried about others’ copying its microcomputer because to do so would require buying the same chips from the same suppliers as IBM, though probably in lower volumes and hence at higher prices than Big Blue was paying. The planners in Boca saw how other companies could copy Acorn, but they did not see how it could be done at a profit.
The only thing remarkable about the IBM Personal Computer was that it was designed and built by IBM. The PC was deliberately positioned against the Apple II. Compared to the Apple II, the PC was big and clunky, but that was by design. It was supposed to look more like a piece of office equipment, while the Apple II was at home on a shelf underneath the family television. The PC had a bigger screen, a larger keyboard on a long cord, and floppy disk drives that held more data (160K each) than did Apple II floppies. Not that anyone could imagine needing 640K of random access memory—that’s how much you could pack into an IBM PC, that is, after some third-party manufacturer came up with a circuit card that held enough memory chips.
The PC looked substantial and had slots, lots of slots for add-in cards. Here was a machine that looked as if it could be expanded forever. Never mind that the anemic power supply was not strong enough to power a PC with all slots filled.
The PC had its husky look, its new operating system (though one that looked reassuringly like CP/M). It had the IBM name and a massive promotion budget to go with it. The new machine even had applications that were provided by third parties but initially released under an IBM label. There was the EasyWriter word processing package and two spreadsheets -- Multiplan, from Microsoft, and VisiCalc.
The PC was a big success and rapidly became the top-selling microcomputer. But it wasn’t a significantly better VisiCalc machine than was the Apple II or its follow-on, the Apple III. VisiCalc and Multiplan looked exactly the same on the screen of an Apple II or an IBM PC, and neither program was significantly faster on the IBM platform either. If it was going to realize its full potential, the PC would still need a compelling application -- one that offered features never before seen on a personal computer and that tied those features specifically to the IBM PC platform so buyers would see no choice but to buy an IBM PC. Every successful computer needs at least one of these compelling applications, remember?
Reprinted with permission