Skillful Death (50 page)

The age of the man behind him was a little easier to peg. The man leaning against the wall had long hair, parted in the middle, which seemed to match the drooping mustache which trailed down either side of his mouth. With the hair, wide tie, and wide collar on his button-down shirt, the man was certainly somewhere between twenty-five and thirty years old.
 

Bud dropped the machine into its monitor mode so he could view and manipulate the memory directly. In that mode, he could see the state of the machine, and enter instructions directly for the machine to execute when he resumed his program. He was speaking directly to the machine, bypassing all the conveniences that most programmers used. The way Bud controlled the machine required a flawless understanding of the workings of the computer.

Bud typed the machine codes of his instructions from memory. While his program executed, Bud made notes about the subroutine he’d just invented.

The man leaning against the wall broke his silence, startling Bud. “You’ve memorized the whole instruction set?”

Bud turned halfway around to get a look at the man.

“Just the parts I use,” he said, with a smile.

“We’ve got an assembler you can use,” the man said. “It’s in that box of disks.” He was referring to a helper application that would help a coder develop a program.

“I’ve never used one,” Bud said. “I wouldn’t know where to start.”

“You can save time. You don’t have to memorize everything, and you can use abstractions to generate multiple machine codes from one line. It compacts everything and makes it simpler.”

“Thanks,” Bud said.
 

The man shrugged when he saw that Bud had no intention of changing his methods.

“I’m Bernie,” the man said. He held out his hand and approached. “Bernie Shaulen.”

“Oh,” Bud said. “Don’t you have a lecture tonight?”

“Yeah, in a few minutes. It’s in the classroom next door. Are you coming?”

“I thought I would work on my program a bit.”

“You’ll get kicked out when the guard comes by. He hates coders. Doesn’t matter if you have a pass or not, he’ll kick you out after eight, every time.”

“Oh,” Bud said. “When does it open again?”

“You’re a student, right?”

“I’m afraid not.”

“That makes it tougher. The lab is only open to students. That’s easy enough to fix. Go over to admissions in the morning and you can enroll. If you’re a resident, you can audit one class per semester for a tiny fee. Just sign up for my class and you’ll get access to the lab from eight until eight.”

“Twelve hours?”

“Every day except holidays.”

♣
   
♢
   
♡
   
♠

Bud gave notice at the repair shop the next morning. He moved out of his apartment and moved into an efficiency near the university. For two months, he spent twelve hours a day in the computer lab. He arrived when the morning guard unlocked the door. He went home sweaty, scruffy, and hungry, twelve hours later. His obsession consumed him.

One morning, his fingers were still. His program pushed the machine to its limits, and he couldn’t think of a way to free space for more instructions. His program was nearly complete, but he had drowned the machine in code.

Bernie glanced through the doorway as he walked down the hall and stopped dead in his tracks. He knew Bud would be there, sitting at the same machine. What made him stop was that he’d never seen Bud sitting and not typing or writing. He approached slowly, waiting to see if Bud would blink or breathe. When Bud turned towards Bernie, the professor nearly jumped back.

“Hey, Bud.”

“Hello, Mr. Bernie.”

“Is there something wrong? Some problem with the machine?” Bernie asked. He looked at the screen and saw a block of memory with high address numbers.

“I’ve run out.”

“Run out?”

“Of space.”

“No, that can’t be,” Bernie said. “These machines are fully upgraded. What could you be doing to fill the whole thing?”

“I was working on my inductor model, trying to make the transients realistic.”

“Your what?”

“I wrote a program to simulate circuits,” Bud said. He broke from the machine monitor and executed his program. The screen filled with a circuit diagram and diagnostic readouts. “You enter your circuit here, and then when you execute, the program simulates all the voltages and currents. You can use it design and test a circuit without having to build it.”

“That’s amazing,” Bernie said, leaning closer. “Does it work?”

“Yes, except the transient isn’t smooth enough. You can graph it here,” Bud said, as he pushed a button. The screen changed to a graph of voltage and current. “You see how coarse that is? I want to smooth it out like this.” Bud switched to a view of a capacitor. The curve was pleasingly smooth.

“Can’t you use the same curve? Aren’t they basically the same curve?”

“One is current, the other voltage,” Bud said.

“Okay, sure, but can’t you just substitute one for the other? Why do you need two different functions?”

“I don’t know how I would re-use the function.”

“Perhaps you should attend my lecture today. We’re talking about parameters and code re-use.”

“I really need to finish this.”

“I think you need to attend my lecture so you can figure out how to finish.”

Bud thought about the book he’d bought in San Francisco. Before that book, he’d always believed that he could only learn by doing. The book had shown him a new way of learning. Perhaps Bernie could teach him in the same way. Bud packed up his stuff and followed the professor into the classroom.

The students looked like miniature versions of Bernie. They all looked familiar since he’d seen each of them in the lab at some point. Bud took a seat near the door and pulled out his notebook. He studied his code while he waited for the lecture to start.

Bernie tailored the lecture for Bud. He gave a contrived example of a routine that had three jobs to do and an absurdly small amount of space in which to execute. First, Bernie tackled just one of the problems, showing how he could design a program in the space to accomplish the task. He repeated the procedure for the other two tasks. The students could see that the three functions would never all fit in memory at the same time.

Bud saw his own problem reflected in the program scrawled on the chalkboard.

Next, Bernie rewrote his solution, feeding a parameter into his routine and allowing that parameter to control the execution. With a clever use of the parameter, Bernie made the routine change its nature. Suddenly, one block of code had multiple meanings depending on the context.
 

The world split apart as Bud’s mind opened to the possibilities. He realized in an instant that he’d been thinking too linearly. His whole approach had been restrained and limited. With Bernie’s method, he could cut the size of his program in half, and have all the room he’d ever need to finish. Bud looked around the room, expecting to see the same wonder he felt reflected on the faces of the other students. They all looked bored. Most weren’t even taking notes on the lecture. Could it be that they already knew this approach? How could they just sit there? Knowing what they knew they should be fighting over each other to get to the lab. Bud forced himself to stay put and hear the rest of the lecture. He was glad he did. He learned several more techniques to optimize his methods.

Bud began writing in his notebook. He didn’t write what Bernie put on the board, but wrote all the ideas which Bernie ignited in his brain.

♣
   
♢
   
♡
   
♠

Bud was living on his modest savings, and they were running low. He ignored his dwindling pile of cash, and focused on finishing his program. As he completed his working circuit simulator, Bud realized that he would never call the program done, but it was certainly useful.

He showed the program to Bernie.

“And you’ve tested it?” Bernie asked.

“Against a hundred circuits,” Bud said. “It works.”

A group of students, sensing that a completed program was being unveiled, gathered around Bernie and Bud.

“Show me one from the beginning.”

Bud restarted the program.

“You move your cursor around the screen like this. You can insert any one of these electronic parts. When you drop a part, it doesn’t have any connections until you hit ‘Enter’ and then draw a line to another part. You can create a node like this.” Bud designed a simple amplifier on the screen. “When I’m done, I simulate like this. Here’s my input voltage over time and here’s the output.”

“Wow,” several students said, as Bud brought up the graphs.

“You see that distortion here?” Bud asked. “If you build a prototype, you’re never sure if the distortion is coming from the long wires, or a bad component, or interference. With this simulator, you can eliminate all those variables. If I increase this resistance and decrease my output a bit, you’ll see that distortion disappear. That way I can pin down the source.”

“That’s amazing, Bud,” Bernie said. “You’ve really got something here. How much are you going to sell it for?”

“Sell it?”

Bernie laughed. “I know you built this for yourself, but surely electronics manufacturers would like to purchase it?”

“Don’t they already have programs like this?”

“Bud, they have programs on mainframes that require a dedicated engineer to feed in all the parts and variables. They can do the same simulation, but it takes a week to get back the results. Then, if they want to tweak something, it’s another week to modify. You can take what you’ve done and put it directly in the hands of the engineers. Think of the cost savings.”

“I wouldn’t know where to start.”

“That’s the great thing about being at a university—we have every discipline right here.”

♣
   
♢
   
♡
   
♠

Bernie and Bud met with the head of Electrical Engineering who set up a demonstration with Sterling Electronics. The CEO of the company wanted to purchase Bud’s program on the spot, but Bernie asked for more time. Bernie practically dragged Bud back to his office and sat him down.

“Bud, this isn’t the type of program you sell,” Bernie said.

“Why not? I have rent to pay. With what they’re offering, I could pay my rent for three years.”

“Because what they’re offering is the same amount that they’d pay to just
use
your program for a year. Think about how much they’re going to save by using your program. With the same number of engineers, they can double their productivity. How much do you think that’s worth? Don’t just think about Sterling Electronics, think about all the electronics companies in California. They’d all pay the same amount or more. If you sell outright to Sterling, then they’re going to license your program to everyone else. They’ll make millions based on your work.”

“And I’ll be happy while I work on the next program.”

“I won’t let you do it. Let’s go back and offer them ten seats. They can run ten copies of your program at a time, for the price they offered. I have a friend who can draw up a license agreement. You’ll see. They’ll jump at the opportunity. If they don’t, we can sweeten the deal by offering them quarterly upgrades. You just add a couple new graph types every few months and they’ll think they’ve won the lottery.”

“Okay,” Bud said.

“What will you call your new company? BudSoft?”

“Dikiware.”

♣
   
♢
   
♡
   
♠

Good news begets good news. After settling the agreement with Sterling, Bud took out a small advertisement in several computer and electronics magazines. He arranged for a post office box, and had to pick up his mail daily to keep up with the requests for information about his program.

Forwarded from his old boss at the repair shop, Bud received a letter from the radio manufacturer. They thanked him for his interest in their circuit, and offered him a contract to use his optimization in their next radio. Bud wrote a quick letter back accepting the deal. He included a pamphlet on his Dikiware circuit analyzer.

The first version of Bud’s software sold moderately well. He sold installations to companies throughout Washington, Oregon, and California. The larger companies relied on their entrenched practices. Those companies utilized departments of engineers to accomplish what Bud’s software simulated, and those men generated dozens of arguments against changing the status quo. Bud didn’t press. For the first time in the U.S., he saw enough income to live well. He simply ignored the large companies, working instead for the younger, hungrier ones.

When his real success came, it happened in a week.
 

Bud sold three seats of his circuit analyzer to a man named Roger Turnbill, who worked for a company that made hearing aids. Their circuits were so small that the physics of the simulation required some tweaks. With their close traces and low voltages, the electronics didn’t behave exactly as Bud’s ideal components. Roger worked directly with Bud to customize the software, and Bud did the work for free. He was anxious to satisfy his customers and always welcomed an opportunity to improve his software.

Roger didn’t last with the hearing aid company though. Just as Bud finished the improvements for him, Roger was hired by a large company who opened a new division. The company made kitchen appliances, but wanted to branch out into handheld electronics—kid’s games. With his experience with miniature circuits, Roger was a natural choice.
 

Roger opened a new plant just south of San Francisco. One of his first calls from his new office was to Bud.

“I need twenty seats, and a boatload of custom components,” Roger said.

“Improving on your hearing aid design?” Bud asked.

“No. I’m starting something new. This is all about integrated circuits.”

Bud caught a flight that afternoon so he could meet with Roger’s new engineers. By the end of the day, he’d sketched out a dozen new components, and a list of features to suit Roger’s needs. The next morning, Bud contracted a public relations firm in Seattle to send out a press release describing the new features of Dikiware. Within two days of the release, orders poured in for the new version. Many of the established manufacturers were struggling with the same problem that Roger managed to articulate .They had to apply their existing workforce to solve the new digital electronic problems. Bud’s software eased that transition magically, and regardless of how high he set the price, they lined up to license it.