Dealers of Lightning (6 page)

In those days ARPA did science, not engineering. Founded in the
nationwide panic that followed the Soviet Union's launch of Sputnik in
1957, the agency at first focused almost exclusively on missile
physics—specifically, how to bring the United States quickly up to par
with the Soviets in shooting projectiles into Earth orbit. By 1962, how­ever, when Licklider was tapped to run a new Information Processing
Techniques Office, or IPTO, the urgency had waned. Whatever mili­tary orientation ARPA still harbored was visible only as a sort of
artifact, as when Licklider discovered by accident that one "cloak-
and-dagger" project under his nominal jurisdiction was so highly classi­fied even he was not cleared to know what it was. ("That made me
nervous," he admitted later.)

ARPA had refocused itself on civilian research in broad scientific
areas, some of them having only tenuous relevance to national security.
"I did not feel much pressure to make a military case for anything,"
Licklider told an interviewer years later. Of course, the Pentagon did
expect that the agency's work might serendipitously lead to solutions of
some of its technical problems, such as the vexing issue of "command
and control": how to employ effectively the immense volume of infor­mation generated on a battlefield to manage the armed forces' increas­ingly elaborate weapons systems.

For years the military had viewed this issue in terms of training the
human beings with their fingers on the triggers. Licklider informed his
bosses that the real solution lay in making the machine meet the human
halfway. This was something he called "man-computer symbiosis," a sub­ject on which he happened to have published a paper two years earlier.
Traditionally, problems had to be written and presented to the computer
very carefully so the machine would understand every step. One tiny
error, and all computation would cease. Try planning a battle under these
conditions—you would be obliterated before reaching the second step in
the process. But what if the system were designed so the computer was
no longer a mute data manipulator, but a participant in a dialogue—
something, he had written in that
paper, like
"a colleague whose compe­tence supplements your own?"

Nothing like that was possible
given the
technology of the time, but it
could be foreseen. "Every time
I
had
the
chance to talk,
I
said
the
mis­sion is interactive computing," Licklider
said. "I
thought, this is going to
revolutionize how people think, how
things
are done."
He
promptly allo­cated most of the money in his budget
to its
pursuit.

Licklider was a tall Midwesterner
whose
owlish glasses camouflaged
a warm and pleasant personality.
He loved
nothing more than bringing
people together and insisted that
even new
acquaintances address him
as "Lick."
A
few months after joining
ARPA
he convened his summit of
government agencies with computer research projects in their bud­gets.
The
group was an august one, encompassing not only
ARPA
and
NASA
but the research arms of
the Navy,
Army, and
Air Force;
the
National
Institutes of Health; and the National Science Foundation.
When
Lick opened the session by describing his own "very modest"
program he left no doubt about who stood at the top of the computer
funding pyramid.
His
$14 million budget, which already supported
projects at
MIT,
Berkeley, and Carnegie-Mellon University, was larger
than those of all the other agencies
combined.

Licklider's program launched the golden age of government-funded
computing research. Very soon he had established a full stable of academic
scholars entitled to come back to his well whenever an appropriate new
line of study struck their fancy. And Licklider defined "appropriate"
broadly.
He
understood that computer research differed from traditional
sciences like physics and chemistry, which lumbered incrementally from
discovery to discovery, building on centuries of theory and experiment in
a
process that resembled geological accretion. Computing, by contrast,
was
young and explosive, driven forward as though by a series of pistol
shots,
eveiy technological innovation inspiring a headlong leap ahead.

His
Information Processing Techniques Office accordingly awarded
its
contracts without any of the bureaucratic paperwork other agencies
required. Recalled
Wes
Clark, who had introduced Licklider to his
first digital computer—the Clark-designed
TX-2—
when they shared
an office floor at
MIT "I
almost felt as though
I
was called up from
time to time to see if I wouldn't be willing to take another quarter of a
million dollars off their hands."

By the time he left IPTO in 1964 to return to MIT, Licklider had set in
motion numerous trailblazing projects aimed at making the computer
more accessible to the user. Studies in graphics pointed toward new ways
of displaying computer-generated information. There were initiatives in
computer networking and new programming languages. Systems to reor­ganize the computer's memory and processing cycles so it might serve
many users simultaneously
—
which was known as time-sharing—
brought the per-session cost of building and running these enormous
contraptions down to a level that even midsized and small universities
could afford.

Lick's successor seemed tire perfect man to manage this expanding
program. Ivan Sutherland was a brilliant MIT graduate who happened to
be serving with the Army as a first lieutenant. Only twenty-six, he had
already amassed an enviable research record, the crowning achievement
of which had been the development of the first interactive computer
graphics program. Known as Sketchpad, the system allowed a user to
draw highly detailed and complex drawings directly on a computer's
cathode-ray screen using a light pen and store them in memory. Among
tire bonds uniting Lick and Sutherland was that the latter had designed
Sketchpad during a protracted series of half-hour sessions on that very
same TX-2 computer. Licklider left Sutherland with a $15 million budget, a workload that
had grown far beyond what a single man could handle, and a suggested
deputy: Bob Taylor. At first glance he was a strange choice. Taylor had
never taken an advanced course in computing. He would never be able
to design hardware or write a software program. But he displayed two
qualities Licklider found appealing: an instinctive grasp of the promise of
man-computer interaction, and an exceptionally high degree of "people
skills." Next to this, Lick and Sutherland figured, his inability to get down
among the bits and electrons was scarcely significant. They could not
have known that within a few short years Bob Taylor would outshine
them both in his influence over computer research in the United States.
The humble job of serving as deputy to a first lieutenant with a Pentagon
staff
appointment was about to
set
Bob Taylor on the path
to
his, and the
computers, destiny.

Ivan
Sutherland spent scarcely
eighteen months
at
IPTO. Late
in 1965
Harvard offered him a tenured position.
He
was officially gone by
June
1966 but unofficially much
earlier. By January
or February that year
Tay­
lor was already running
IPTO all
by himself.

The
apprenticeship had
been
short,
but
edifying enough. Taylor
rapidly digested such important
conventions of
Pentagon life as the enor­mous significance of tiny gradations in
rank.
One day he learned drat his
own
new assistant, a twenty-three-year-old
MIT
engineer
named Barry
Wessler,
was to be assigned
the
humble
Chilian
rank of
GS-9. Too
low:
Taylor concluded the appropriate level
would
be GS-13.
The
difference
was a trivial one in terms of
pay,
which
was
nearly the same for both
grades. But Taylor reasoned that the higher rank would guarantee that
Wessler got treated by the
Pentagon’s brass
hats as a seasoned profes­sional rather than
a
greenhorn. This
was a
distinction certain to return
dividends day after day in terms
of Wessler's
interaction with military offi­cialdom. (Plus there would be
the satisfaction
of having squeezed
Penta­
gon rules until they squealed.)

As
Wessler recalled, "Bob worked
by finding
out what
he
couldn't
do,
and then going for it.
He
could
have
gotten me a level or two
higher without too much trouble, but
GS-13?
That was off the scale.
But
he got it done."