Dealers of Lightning (40 page)

Metcalfe himself did not realize the extent to which his offspring had
become a indispensable part of PARC's lifestyle until one day shortly
after EARS was launched. After accidentally disabling the ether by
removing a piece of hardware he noticed "one after another of my col­leagues popping up, wondering why the network was down." Sneakernet,
obviously, was dead.

Two more important events happened that same year. On March 31
Metcalfe filed for a patent on Ethernet in his own name and those of
Boggs, Thacker, and Lampson, each of whom had contributed a critical
element of the technology. It was awarded (and assigned to Xerox) two
years later.

Second, he resigned. Metcalfe had worked at PARC for three years,
about as long as he had expected to. Industry headhunters were calling.
"I was contemplating moving on and I was also contemplating staying,"
he recalled. Before deciding he sat down with his immediate boss, Jerry
Elkind.

"What would make you stay?" Elkind asked.

"I said, Well, Jerry, if I did stay, how long would it take me to get
your job?'
"
Metcalfe recollected. "Elkind thought about it. He said,
'Well, gee, you're a member of the research staff, and then you'll be a
senior member of the research staff, and then you'll become a consult­ing member of the research staff'. .
.
And basically Jerry Elkind, who
was twenty years older than me, said it would take me about twenty
years to get his job. I told him that wasn't what I had in mind."

Anyway, the decision had already been forced upon him. His wife of
seven years asked for a divorce. Simultaneously he was offered a job in
Los Angeles by Citibank, which was planning to redeploy its aging elec­tronic fund transfer system onto new computers. "So here's this job at
Citibank where I'm to get a thirty per cent raise and an office with a view
of Catalina Island and a chance to live in L.A., which was appealing at the
time. So I bolted."

Metcalfe's departure rattled the PARC staff like a tremor on the San
Andreas fault. This was not only because he was the first top computer
scientist to quit PARC since its founding five years earlier. More impor­tant, his resignation provided the first hint that while they had buried
themselves in their research Camelot, a whole new world had sprung up
outside—and that it would welcome them and their knowledge.

What they could not know was that in a very short time Metcalfe would be back.

This is the story of how three sheets of lined yellow
paper and three misfits spawned an industry.

The
industry is desktop publishing, which today
allows millions of ordinary persons to turn out newsletters, magazines,
and books as though they were professionals; and allows millions of
professionals—writers, editors, and publishers
—
to do their even more
sophisticated work faster and easier.

The three sheets of yellow paper belonged to Butler Lampson.
Cov­
ered with functions and algorithms written in his neat, angular hand, they
represented his first pass at designing a text editor
—
a word processing
program, if you will—for the Alto, which at that point did not have one.

The three misfits were Charles Simonyi, who one day happened upon
Lampson’s scribbles and asked what they meant; and Larry Tesler and
Tim Mott, who had been assigned to help
Xerox's
textbook subsidiary
find a way to make the editing and rewriting of manuscripts somewhat
less tedious and time-consuming than the laying of bricks.

Together they achieved what Lampson later termed "one of the most
successful collaborations in the history of
PARC"
(and that is a history
rife with spectacular partnerships). The trio gave the Alto its "killer
app"—the application that burned its unique virtues indelibly into
people’s minds—as well as the program that first showed professionals
outside PARC how the personal computer might improve their lives.

These programs were called Bravo and Gypsy. Their development
consumed more than three years, starting with the moment when
Simonyi reached out his hand for Lampson s three yellow pages and
said, "Can I take a look?"

With his mop of straight brown hair and his deep-set blue eyes,
Simonyi might have stepped directly out of a Jacques-Louis David por­trait of the young Napoleon. The similarity did not end there. Like
Napoleon, Simonyi was a young man who rarely lacked the self-
assurance to tell his elders where they were wrong and he was right.
They also both came to prominence as outsiders: As Napoleon had
come to the French revolutionary army from the rugged Mediter­ranean island of Corsica, Simonyi had crossed to the United States
from the socialist hell of 1960s Hungary.

The elder son of a Budapest professor of electrical engineering,
Simonyi had first encountered a computer at the age of fifteen. This was
a Soviet-made contraption called the Ural II. The Ural was one of five
computers in the whole country and a time machine of a unique variety.
"All the action on this computer was directed through the console—it
was truly a hands-on, one-on-one experience," Simonyi recalled. "It was
exactly like the personal computer of fifteen years later, because it was
just you and the machine and no one else."

The Ural could have been the model for the computer in every science
fiction film of the 1950s and 1960s. The size of a large room, it was driven
by thousands of vacuum tubes glowing with an eerie orange light. The
operators console was like the keyboard of an old-fashioned cash regis­ter—six columns of numbered switches and an
enter
key on the right, all
operated by a mechanism with substantial Soviet heft. "All this was very
exhilarating because there was a lot of noise associated with it," Simonyi
recalled. "Every time I hit the switch it clicked very firmly. Whenever I
cleared it, all the keys released at once with a great
'Thunk!'"

Housed at a Budapest engineering institute, the Ural bedeviled its
operators by blowing out at least one tube every time it was switched on.
The only remedy was never to turn it off, which meant hiring someone to
babysit the behemoth all night after everyone went home. Through his
father, Simonyi wangled the job for himself. With the help of a mentor on
the university faculty and the endless, empty nights available for full-time
experimentation, he soon taught himself all there was to know about pro­gramming in Octal, the base-8 system on which the Ural was pro­grammed. His first programs were designed to fill in "magic squares,"
giant grids of numbers in which all the columns and rows add up to the
same sum. Years later he could still remember how he would spend hours
punching buttons on the machine to create magic squares eighty cells
wide by eighty cells deep, then arrive home in the morning "with an
incredible headache and giant rolls of paper printouts."

After about a year a Danish computer technician he met at a Budapest
trade fair offered him a job. Simonyi was sixteen. All that prevented him
from leaving the country on a temporary pass was the Hungarian mili­tary's craving for draftees. "The way I went around them was that I was
underage, so they couldn't draft me, and if you were in college you got a
deferment. So I got myself admitted to the university and told them if
you don't let me go to Denmark I'll go to university and you won't get me.
If you let me go, I'm coming back in one year, and
then
you'll have me."

Of course there was no question of his going back. This was 1966.
Everywhere in Hungary memories of the aborted revolution of 1956
were still painfully fresh. "One thing that shocked me when I got to Den­mark," he recollected, "was that the houses didn't have bullet holes in
them." Back home his father lost his job in retaliation for his son's defec­tion. "But he had a lot of political problems anyway, so this was just on top
of it. Plus it was calculated in—either that I would have an unhappy life
or he would have one more political problem. My parents agreed. My
dad was practically pushing me to go."

About a year and a half later Simonyi left Denmark for the United
States. Because his situation at home was deemed not to have been per­ilous enough to warrant political status, he arrived on a student visa,
which prohibited employment. "So I told the authorities that due to
extraordinary circumstances I had to take up work. The extraordinary
circumstances were that I was running out of money." The job he found
was at Berkeley Computer, where he encountered the troika of Thacker,
Lampson, and Taylor and survived the one and only corporate bank­ruptcy of his life.

Following BCCs collapse, Simonyi had tagged along with Mel Pirtle
to his next job, which was to supervise the building of the Illiac IV on
an ARPA contract at Ames Research Center, a NASA facility just south
of Palo Alto. Illiac was a vast, overdesigned attempt at a large-scale
system that some called the first supercomputer and others called
computing’s Vietnam. (It never became fully functional, despite the
expenditure of millions of dollars.) Simonyi tended toward the latter
view, which is not to say he found the program entirely worthless.
Aside from what it taught him about computer architecture, as a gov­ernment program Illiac at one point saved him from being perma­nently evicted from the United States. This occurred when he left the
country for a couple of days one January to visit West Germany—his
first chance in seven or eight years to see his father, who was giving a
lecture in Hamburg. In his excitement he overlooked not only his
overcoat but the rule that once the holder of a student visa leaves the
country he must reapply for permission to come back.

"So Friday evening at four o'clock I went to the embassy to say, 'Hey,
my plane's leaving in one hour, would you please give me a visa.' They
took one look at me and said no. I called Pirtle right away and the
wheels started to turn. They opened the embassy on Saturday, just to
give me the visa stamp."

Ever since his first labor-intensive experience with the Ural, Simonyi
had been fascinated by the art of programming. By 1972, when he
rejoined his BCC mates at PARC, he had come up with a less trying
methodology, which he labeled "meta-programming" and made the
topic of his Stanford doctoral thesis. Meta-programming involved a
team leader's drafting a detailed blueprint for a program using a highly
abstract language, and handing this over to assistants for the actual
coding of the software. The idea, as Simonyi described it with his char­acteristic bluntness, was to improve everybody's productivity by giving
the smartest programmer the freedom to think in broad strokes while
a couple of overworked assistants reduced his ideas to code that the
machine could understand. In essence, Simonyi was programming the
programmers.

His first experiment in the process, which involved hiring two under­graduates from Stanford as these intellectual menials, he called Alpha.
Around the time he was ready to conduct a second experiment, he found
himself in Lampson's office, studying the three canary-colored pages.

"What is this?" he asked.

"We need a text editor for the Alto," Lampson replied. "Nobody's
working on it, so I thought I'd start."

Lampson was being slightly disingenuous. The Alto did not just need a
text editor—it needed
everything.
The machine had been around for
nearly a year and, quite clearly, the novelty of Cookie Monster had begun
to wear off. "Some people didn't really see the potential of Alto," Lampson recalled later. "We were trying to draw more people into it, because
obviously the thing is useless without software. For the first year or so
after it existed it wasn't very interesting because it didn't have very inter­esting software." The yellow sheets were Lampson's way of jump-starting
the process.

Highly intrigued, Simonyi ran his eyes over Lampson's formulas. He
fancied himself a great programmer, which he was, having learned the art
on one of the most recalcitrant computers ever built. But he was also
aware, as he said later, that "it's not enough to be a great programmer; you
have to find a great problem."

This looked like such a problem. Editing text on a graphical screen
seemed easy at first glance, but it was rife with hidden difficulties and
unexplored potential.

"I thought we were on the cusp of a paradigm shift," he said later. "I
could see books in their entirety flowing in front of you, virtual books and
everything. In retrospect it seems so obvious. Uh-uh, it wasn't obvious to
anyone. This stuff was in the future then. But it was suddenly clear to me
that with the combination of Xerox and this machine, word processing
was going to be a key application. I took it and decided to make it happen,
because it looked very sweet." Since it would be the second experiment
undertaken for his doctorate Simonyi
moved
one step down in the alpha­
bet, and
called it Bravo.