(FORTUNE Magazine) – It seemed like a great idea at the time. In February 1987, during a cross-country flight in a private jet, Mitch Kapor, the founder of Lotus Development Corp., and Jerry Kaplan, an artificial intelligence wizard, hatched the idea of a hand-held, pen-operated computer. It would weigh little more than an Etch-A-Sketch but be able to read and manipulate handwritten letters, numbers, and drawings, and pack the smarts of a high-powered PC.

That brainstorm at 35,000 feet launched Kaplan in the game that has become the hallmark of U.S. capitalism: founding a company, attracting venture capital, and trying to stake out a new market before the like of IBM, Microsoft, and AT&T claim it as their own. Kaplan and his associates set out to change the way people work; in a new book, he tells how he learned instead what can happen to a young company when its timing is wrong.

The startup game is an elaborate contest created to accelerate the pace at which corporations evolve, played continuously by an endless parade of hopeful entrepreneurs. It is a carnival game of life, testing the strength, aim, and skill of contestants willing to expose themselves, for glory or ridicule, to public scrutiny. The challenge is to find a new or better way to do business; the rewards are increased wealth, enhanced personal reputation, and control over one's own destiny. The startup game is designed to motivate our brightest, most creative, and hardest-working individuals to improve the use of society's resources, increase employment, and provide a broader range of quality goods and services. Here's how the game is played.

It begins with an aspiring entrepreneur who is willing to step right up and be tested. As in many other games, the player starts with an artificial currency--in this case, the stock of the new venture. The goal is simple: Increase the value of the entrepreneur's shares, because when the game is over, these can be cashed in for real money. The trick is to swap some of the stock for three resources--ideas, money, and people--then use these resources to increase the value of the remaining stock.

The entrepreneur doesn't raise all the required money up front because that would mean selling too much of the stock. Usually, the initial investment is just enough to reach some identifiable milestone. This milestone is chosen to demonstrate to potential future investors that the company's prospects have improved, justifying a higher price for the stock, so that less of it has to be sold. If the money runs out before the milestone is reached, the game is over. In the meantime, other companies may try to steal the ideas and the people, and may even try to run the venture out of money.

Venture capitalists typically look for a reasonable possibility of making five to ten times their original investment within five years. This corresponds to borrowing money at 50% per year or more, a cost of capital that makes the startup game a race against the clock.

The final step in the game is a financing event called an initial public offering, or IPO. The IPO--when the company is first listed on a public stock exchange--usually marks the transition of the enterprise from a risky venture to a profitable company. Up until this time, it is virtually impossible to dispose of stock and get cash. Soon after the IPO, the entrepreneur is free to sell stock on the open market, as are the investors and employees. This marks the successful end of the startup game and the creation of a viable company. The entrepreneur can now cash in his stock and go home, take on the less risky job of managing a going concern, or step up and play the game again.

In June 1987, Mitchell Kapor arranged for me an introduction to the legendary venture capitalist John Doerr, whose firm of Kleiner Perkins Caufield & Byers had backed such prominent technology companies as Tandem, Compaq, Sun, and Lotus. John asked me to drop by his office the following Monday to chat with some of his partners. I mistook this for an informal bull session, but any experienced entrepreneur would have known better.

To VCs, Monday is as sacred as Sunday is to the Vatican. This is the day when venture capital partnerships around the world have their official meetings to review potential investments. All the partners participate, come hell or high water--by videoconference call from a branch office if possible, or by speaker phone from their deathbed if necessary. An invitation to address the partners at a top-tier firm like Kleiner Perkins--or KP, as it is known in the trade--is an unusual opportunity. I was unprepared.

At the appointed hour, I showed up wearing a sport jacket, with my shirt open at the collar. I carried little else but a maroon leather portfolio holding a tablet of paper and a pen that I had received as a Christmas gift. No business plan, no 35-mm slides, no charts, no financial projections, no prototypes.

The Kleiner Perkins offices were on the 35th floor of a posh office tower in the heart of San Francisco's financial district. The floor-to-ceiling windows framed a spectacular panoramic view of the bay on one side and the city on the other. The partners' offices were extravagant by most standards, set apart from each other and from a large common area by smoked-glass partitions.

The previous presenter was just finishing up. Dressed in a dark blue pinstriped suit, his red power tie thrust forward with a gold tie pin, he was nervously fielding rapid-fire questions from a brigade of partners and associates inexplicably packed into a small corner conference room. A bare prototype circuitboard sat on the conference table. A crisp color graph was projected on a whiteboard; the man's perspiration gleamed in the reflected blue light.

"Thank you. We'll let you know our decision in about a week," said one of the partners. The presenter collected his belongings and left quietly.

After a short break, John Doerr reassembled the group and invited me in. He made a short introduction, covering my background and explaining Mitchell Kapor's interest in the project, then turned the meeting over to me. I was nearing a state of panic. I paused to size things up, knowing that the brief silence before diving in would create a momentary impression of authority. In reality, I was searching for a strategy.

I had a flash of deja vu. I remembered facing this same kind of "show me" crowd at my Ph.D. dissertation defense. The key to success there was in recognizing that although the examination committee had the power, I had the knowledge. The same was true here. I had done a lot more thinking about this topic than anyone else in the room, and to win their respect, I merely had to demonstrate this fact by keeping the discussion focused on the areas I had already investigated. I decided to lead with the business issues.

"Gentlemen, you probably think that there is no longer any way to make money by starting a new personal computer company. The competition is brutal, and the barriers to entry are high. However, I'm here to suggest to you the possibility that the PC as we know and love it may not be the best and final form that computers will take. I believe that a new type of computer, more like a notebook than a typewriter and operated by a pen rather than a keyboard, will serve the needs of professionals like ourselves when we are away from our desks. We will use them to take notes; send and receive messages through cellular telephone links; look up addresses, phone numbers, price lists, and inventories; do spreadsheet calculations; and fill out order forms. All of this can be done unobtrusively while sitting in meetings, conferring with clients, commuting to work on the train, or even when standing up and walking around. Like the fax machine, the pen computer can dramatically accelerate the pace and increase the efficiency at which business can be conducted.

"I can't say for sure when this will occur. But I do know that it will happen, and someone is going to make a big pile of money on it. With a little luck and some hard work, I think we could be the ones. Like the PC, I think this concept could come out of nowhere and take the industry by storm. I want to be the one to do it."

I proceeded to talk about the required technologies and the state of the art. The plain fact was that the project was technically very risky, and there was no point in hiding it. The greatest risk was whether a machine could reliably recognize handwriting and convert it into its equivalent in computer text, known by the strange acronym ASCII.

My audience seemed tense. I couldn't tell whether they were annoyed by my lack of preparation or merely concentrating on what I was saying. Several people narrowed their eyes disapprovingly--or perhaps they were just deep in thought. I had been talking nonstop for about ten minutes, and figured I'd better close. Thinking I had already blown it, and therefore had little to lose, I decided to risk some theatrics.

"If I were carrying a portable PC right now, you would sure as hell know it. You probably didn't realize that I am holding a model of the future of computing right here in my hand."

I tossed my maroon leather case in the air. It sailed to the center of the table, where it landed with a loud clap.

"Gentlemen, here is a model of the next step in the computer revolution."

For a moment, I thought this final act of drama might get me thrown out of the room. They were sitting in stunned silence, staring at my plain leather folder--which lay motionless on the table--as though it were suddenly going to come to life. Brook Byers, the youthful-looking but longtime partner in the firm, slowly reached out and touched the portfolio as if it were some sort of talisman. He asked the first question.

"Just how much information could you store in something like this?"

John Doerr answered before I could respond. "It doesn't matter. Memory chips are getting smaller and cheaper each year, and the capacity will probably double for the same size and price annually."

Someone else chimed in. "But bear in mind, John, that unless you translate the handwriting into ASCII, it's likely to take up a lot more room." The speaker was Vinod Khosla, a young man born in India and educated at the Stanford business school who was the founding CEO of Sun Microsystems. He acted as a consultant, helping the partnership evaluate technology deals. His cool manner masked a fierce analytical mind and the competitive instincts of a gladiator.

Before John could respond, Frank Caufield, a partner known for his outspoken manner, pointed at him. "Storage won't matter with John's handwriting--it's completely illegible!" This was the break I needed.

"Then you won't mind if I use his writing as a benchmark of our progress," I shot back. The rest of the group fell silent and looked toward Doerr and Byers to gauge their reaction to my impertinence.

They both chuckled, signaling their approval. Then the entire room exploded with laughter. "At least I know how to type," John said.

From that point on, I hardly had to speak. The partners and associates traded good-natured barbs and insights as we fleshed out this new business opportunity. The single conversation split into two threads, then split again, as everyone offered a question or an opinion. It was beginning to sound like a lively cocktail party. Periodically someone reached out to touch or examine my portfolio, which was bouncing back and forth among the participants. It had been magically transformed from a stationery-store accessory into a symbol of the future of technology.

This din continued for several more minutes, after which Brook Byers called the meeting back to order.

"One last question," Byers said. "What are your personal goals for this project?"

I spoke slowly. "Well, I'm not in it to prove that I'm smart or that I can run a big company." I knew these to be two of the most common mistakes that entrepreneurs make. "I guess I have four goals. First, to produce a product that delivers real value to real people. Second, to provide an above-average rate of return for the investors. Third, to create a healthy and challenging work environment for the employees." I paused.

"And fourth?" Byers said, raising his eyebrows.

I couldn't think of a final point. I glanced around and noticed a plate of sandwiches and cookies left over from their working lunch. "To never pay for a meal for four years."

He laughed. "Have some. We're a full-service venture capital firm."

"Thanks." I picked up a cookie.

Then they took a break. Several people scrambled for control of the conference room telephone. The losers bolted out the door in search of another line. John Doerr turned to me and said, "Good job. I'll talk this over with the partners and let you know where we stand."

I thanked him, recovered my portfolio from two associates who were inspecting it closely, and left.

GO decides that its best shot at success is to adopt the strategy Microsoft used with the PC: develop software that will serve as the heart of the machine, and create a standard by licensing it to dozens of hardware manufacturers. The first notepad computer, built by IBM with software by GO, is a smash when the companies demonstrate it to industry leaders at a conference in March 1991.

But all is not well. GO finds itself at the mercy of the large partners on which it depends for hardware and vulnerable to major competitors in software. Microsoft is particularly worrisome. After exploring the possibility of an alliance with GO, Microsoft opts instead to compete. At the same industry conference, it unveils Pen Windows software, which Kaplan and his associates find unnervingly similar to their own product, Penpoint.

The GO officials are dismayed to learn that the same Microsoft technology manager who had visited GO years earlier for a confidential briefing on Penpoint now heads the Microsoft project. For its aggressive tactics in the market, Microsoft is coincidentally coming under antitrust scrutiny; the Federal Trade Commission is summoning rivals to Washington, D.C., to aid in its inquiry.

We all harbor a desperate desire to believe that the world is ultimately predictable. That's why business schools teach formal techniques for decision analysis. But anyone who has managed a startup knows that predictability is an illusion. In this environment, you are faced with an endless stream of arbitrary challenges that bear down on you with the relentlessness of an automatic pitching machine. There's no time to think. The trick is to know when to swing and when to duck. But I suspect that each arbitrary decision is like smoking a pack of cigarettes--somehow, somewhere, it shortens your life. When I saw Microsoft's product and thought about our earlier discussions with them, I wondered if I'd swung when I should have ducked. Exactly four weeks after the demos at the industry conference, John Croll, our in-house attorney, bounded into my office. A short man with bright eyes and a blooming bald spot, he has the temperament of a bantamweight prizefighter. He was winded from running down the hall.

"We got the call!" he shouted.

"What call?"

"The FTC. They told me they're looking into allegations that Microsoft is using its dominant position in operating systems to take over the applications business."

"John, sit down," I said. "You're hyperventilating." But he was too excited to pay attention. Corporate lawyers spend most of their time on boring matters like letters of intent, licenses, and contracts, and this was a chance to do something exciting.

He reluctantly took a seat, then popped up again. "Anyway, they want us to fly to Washington and testify next week--in secret."

From then on, John spent virtually all his waking hours preparing for the trip. He collected every shred of documentation that might bear on the matter--e-mail messages, engineers' notebooks, correspondence--and organized them into a carefully indexed chronological account of our dealings with Microsoft.

Two days before our trip, I was sitting in a meeting with a senior executive of one of the smaller Korean computer manufacturers. This was to be a routine dog-and-pony, where we would give him a demo and pitch him on building a pen computer that could run Penpoint. But I learned that this company was different from the other hardware manufacturers who were parading through our offices two or three times a week: It was not yet in the business of building Microsoft DOS-based PCs. When the conversation turned to Microsoft, the executive grew more animated.

"I don't want to do business with them," he said matter-of-factly.

I was surprised, to say the least. "Why is that?"

"I don't like their pricing arrangements. To get the best price on their operating systems--one that would allow me to compete with other manufacturers--they want me to pay a royalty on every Intel-compatible machine that I ship, whether or not I put their software on it."

He looked so angry, I wasn't sure I understood him correctly. "You mean if you license Penpoint from us and put it on a computer that contains an Intel chip, you would have to pay Microsoft a license fee as well as paying us, even though you were only selling our software?"

"That's correct. So I don't want to do it. That's why I'm here, to see if you have an alternative I can use."

"But that's outrageous." Microsoft had such strong control over its operating-system customers that it had, in effect, found a way to levy a tax on Penpoint, which went right into their pocket. "Look, we've pitched Penpoint to lots of manufacturers, and no one has mentioned this to us."

"That's because they probably already have such a license with Microsoft, and the deal requires them to keep the pricing structure confidential."

Suddenly everything fell into place. I had recently seen one of our licensees' price lists for its upcoming pen computer product, and the company had inexplicably priced Penpoint at almost twice as much as Pen Windows. It also helped explain why, of the many manufacturers that had expressed a high degree of interest in Penpoint, only three had signed a license with us.

In an incredible coincidence, my assistant stuck her head in the room as I was pondering this revelation. "Sorry to interrupt, but there's a call I think you might want to take. It's Bill Gates."

I thanked the visiting executive for coming and excused myself from the meeting. On the way back to my office to take the call, I collected John Croll so he could listen in. "Be careful what you say," John said. "It can be used against you later, in court."

Gates sounded distraught. "I just want you to know that I've seen your comments in the papers. If you think we've done anything wrong, I'd be more than happy to provide whatever information you want on Pen Windows. I don't want to see this thing played out in the press."

I was noncommittal. "Bill, I understand your concern about comments in the press, but you know what it's like. They love to build things up."

"I'd be happy to personally come down and go over this with you if you think there is any problem at all."

"I appreciate the offer, but let me think about it. What I'll do is instruct the staff not to comment to reporters. We're simply considering our options. We aren't going to go off half-cocked and file a lawsuit without looking at the matter very seriously."

"I just hope you understand what you could be getting into. You know our 'look and feel' dispute with Apple?" Gates was referring to one of the longest-running intellectual-property battles in the history of the industry. Back when Apple had first developed the Macintosh, with its graphical user interface, Microsoft was one of the earliest companies to agree to build badly needed applications. But before the applications were released, it was rumored that Gates had persuaded John Sculley to grant him a license to adopt the Macintosh's graphical look by suggesting he might withhold the products. Sculley gave in, thinking there was little that Gates could do with the rights. Years later--after Microsoft introduced Windows, which mimicked many elements of the Macintosh--Sculley realized the seriousness of the threat resulting from this decision and filed suit, claiming that the license covered only an earlier version of Windows. Apple ultimately lost the suit. Now Gates spoke with restrained emphasis. "I've spent over $4 million so far on that one lawsuit!"

He had couched his argument as friendly advice, but we both knew what he was talking about. There was no way a small startup company like GO could possibly afford the cost of legal action against Microsoft. Compared with them, we could barely afford the carfare to the courthouse. A lawsuit would tie us up forever with discovery requests, endless depositions, and baseless counterclaims. If we so much as threatened to sue, it would rain lawyers on our office like the plagues of Egypt.

"That much, huh?" I wasn't going to get into an argument with him, so I changed the subject. "Maybe the first step is for you to allow us to look over a copy of Pen Windows in some detail, if you're willing."

"I'll have one sent down right away."

After we hung up, John narrowed his eyes suspiciously. "He probably got wind that we were about to visit the FTC."

I said, "He did seem to be on his best behavior." Fast-forward two years to 1993. GO by now has 200 employees, its software is on the market, forecasts call for the company to hit several million dollars in annual sales and narrow its losses. Most important, it has found powerful allies. AT&T has invested $10 million; State Farm, the mammoth mutual insurance company, is working on a plan involving the purchase of thousands of pen computers. But GO, eating up cash at a rate of $2 million a month, is running out of time. The strain is heaviest on CEO Bill Campbell, a tough, charismatic industry veteran who had signed on two years earlier and grown passionately committed to both the project and the people.

By early spring, it was evident that the sales projections for 1993 were a pipe dream. Virtually all of the sales had been slated for the fourth quarter, which had now slipped into 1994. Our largest customer, State Farm, which wanted to distribute pen computers to its claims adjusters for use in the field, called to say it was putting the project on hold for a year. As a result, the new $10 million investment from AT&T would stretch only until September 1993.

IBM had already announced a pen computer called the Thinkpad. Now, to our surprise, the company formed yet another task force to determine its strategy for pen computers. It also destroyed what little brand identification it had managed to create around the Thinkpad by co-opting the name for its new line of laptops.

Meanwhile, the euphoria of the publicity for Penpoint had faded considerably, particularly after another startup, General Magic, had pushed the limits of hype into uncharted territory. In early February it staged a major press announcement with big-name corporate partners--AT&T, Sony, Motorola, Matsushita, and Phillips--about its own pen computers. But it showed no product, and Marc Porat, the company's visionary CEO, made the strategic mistake of predicting that it could take ten or 20 years for this technology to have a real impact. For reporters who lived and died by daily deadlines, this was tantamount to promising salvation in the afterlife.

Not all our problems were external, however. GO's product schedules had slipped as well, raising the ominous possibility that we would need to seek out even more money, at a time when we had nothing to show but a bagful of promises and missed schedules.

What was supposed to be a midyear release for the next version of Penpoint--which was to occupy substantially less memory and be much easier to use--began to slide inexorably toward the dead of winter. Most of the engineers were occupied with simplifying portions of the program to reduce its size. But in the course of rewriting, everyone was proposing improvements that would tend to increase its size. Penpoint was succumbing to a disease known as feature creep--the irresistible temptation for engineers to load a product down with their favorite special features. Before unnecessary additions hardened around their feet like concrete, the senior technical managers staged meetings called Feature Court, with "Judge Wapner" presiding. Each side would argue its case, and a binding decision was made on the spot. Despite these efforts, the system seemed to grow like a cancer. No sooner were we able to save some space than a new, indispensable requirement would emerge.

As the depth of these problems became increasingly clear, Bill Campbell reacted by stepping up his already breakneck pace. He spun faster and faster, dashing from meeting to airport to conference call to presentation, until he seemed ready to drop. But what seemed to finally break his spirit was an incident with Compaq Computer Corp., a major manufacturer of IBM-compatible personal computers, which had decided to get serious about selling a pen computer. Our salespeople worked for months to educate Compaq's executives on the market and our product, bidding against Microsoft for the business. After a formal proposal and presentation to senior Compaq officials, they let us know informally that we were the leading contender. Suddenly our staff's phone calls to Compaq went unreturned. Bill learned shortly afterward that Microsoft countered by proposing, allegedly to Compaq's CEO, a closer relationship with Microsoft than that of other personal computer makers--but this arrangement was to include the pen project as well. Compaq had agreed.

On March 1, Bill requested an emergency meeting in his office at six with Robert Carr, our vice president of software development, Randy Komisar, our CFO, and me. Somehow we all knew he had something important to discuss.

Bill looked like hell. His eyes were bloodshot and his head shook periodically to ward off the desire to fall asleep. "Sit down, gentlemen," he said solemnly. "I've come to a difficult conclusion. Since last August, we've been living a lie. There's no way we're going to pull this through. We're spending over $2 million a month with no revenue in sight. Everyone's stretched to their limit." As he spoke, he flexed a rubber band between his fingers ever more vigorously. "I want you guys to know that I don't care what happens to me. The important thing is to save the project and the organization--to protect what we've built."

While Bill often emphasized his points with strong language, he was doing just the opposite now. His restrained wording carried special weight, especially in this private setting. We weren't sure where he was heading, but it didn't sound good. "Robert, Jerry, I think we're going to have to put the company up for adoption--to find it a good home."

We were momentarily stunned into silence. "You're saying we should sell the company," I offered.

He couldn't bring himself to use the words, so he just nodded.

Robert and I looked at each other. Robert's eyes said it all. For years, he had been the trouper, ready to go just one more mile. But this time he looked weary in a way he never had before, as though he wanted to say "enough is enough." Perhaps under the wing of a larger, influential company with a long-term perspective and deep pockets, we could rekindle the sense of optimism and purpose that was so essential.

I spoke for both of us. "We won't be a problem. Do what you have to do. We'll back you up all the way."

Bill turned to Randy. "How about you?"

For once, there wasn't a hint of a smile on Randy's face. He spoke slowly and philosophically. "You know what? The problem is simple. We have a great product that doesn't sell. It seems to me that GO is like Gone with the Wind, a labor of love with no real possibility of a return. Did you know that Selznick was so obsessed with making the movie that he gave away the first seven years of its profits to another studio just to get Clark Gable? We've built something really special here, and I agree with you guys--the first priority is to preserve the project and our people. I can survive losing my job, but it'd kill me if after all the hard work, brilliant engineering, and money, the product died before it had a real shot in the marketplace."

"We're all in agreement, then," Bill said. "I'll talk to the other executive staff. Until I say so, everything's business as usual, okay?" Everyone nodded. "Okay, then. We need some runway while I work up some candidate buyers." An obvious candidate was AT&T. But others, including the government of Taiwan, had expressed strong interest in investing. For once, we had excellent prospects for future financing.

In light of this, it seemed peculiar that we were talking about selling off the company. But as I looked around the room, I realized for the first time that money wasn't the problem. It was a loss of faith: Randy's misgivings about the viability of our business; Robert's doubts about our development schedules; Bill's skepticism that we could develop the market momentum required to succeed as an independent company; and my concern that the management team might not stick it out for however long it took. Despite the expanding bubble of hype about portable computers and mobile communications, our dream seemed as elusive and distant now as it ever had.

Today, GO is gone--AT&T absorbed it and later decided to shut down the project. Recently State Farm completed field tests on an IBM Penpoint machine and placed a $50 million order.

None of GO's executives or engineers saw any money from their stock in the company. Kleiner Perkins recovered a fraction of its multimillion-dollar investment in pen computing through subsequent dealings with AT&T.

Jerry Kaplan recently launched another startup--an online entertainment and shopping service that he says is "intended to be to the Internet what QVC is to television." His Internet address is kaplanj@hmco.com