The Impatient Genius behind Mathematica
Jonathan Littman, MacWEEK 2 (July 12, 1988) 38, 40. Traditionally, math software has found few users. But math software hasn't been designed before by Wolfram.
SANTA CLARA, Calif.—Like most geniuses, Stephen Wolfram usually knows the correct answer before everybody else. This can cause problems. He knew, for instance, that Apple should bundle the Mac version of Mathematica, his revolutionary new software that experts say may transform the field of mathematics.
Last year, Apple officials showed the oft-Birkenstock-shod, bearded, bespectacled physicist a catalog of the Mac's engineering programs. Wolfram, 28, remembered what he considered to be largely a collection of poorly ported programs from the IBM PC: "It was pathetic," he groaned.
No one ever said it was easy knowing all the answers. Wolfram knew Apple needed strong engineering tools like his software, but he also knew that it would be difficult to convince Apple executives to make an exception to the company practice of not bundling software with the Mac.
Wolfram was right. Apple decided not to bundle Mathematica. Instead, Wolfram said, company officials advised him to try to sell the Mac version of his program through Claris. But, as Wolfram recalled it, Claris only seemed interested in selling "generic software tools." Wolfram, a world-renowned physicist, said he tried to be patient. "This is not a word processor, a spreadsheet or a database," he told Claris representatives in an attempt to set up a meeting. "It's unique."
A meeting was arranged and then, according to Wolfram, canceled by Claris the day before it was to take place. "They had never seen my product," said the surprised and offended Wolfram. "They didn't understand it at all." Claris made overtures later, Wolfram said, but it was too late.
A Claris spokesman declined to dispute Wolfram's story and would only say that Mathematica would have been a "radical departure from our general business-application portfolio."
But there is no disputing the extraordinary qualities of Mathematica, or of its creator. Wolfram has seldom waited, often antagonized and almost never done what was expected of him. Born in Britain, Wolfram attended Eton and Oxford but didn't bother to graduate from either lofty institution. Instead, he published his first scientific paper at age 16, and at 17 had his first job as a physicist. Three years later, he received a doctorate from the California Institute of Technology in Pasadena, Calif., and was happy to have skipped over the basic requirements at Eton, like Latin and Greek. Said Wolfram: "Generally, it's considered desirable if one can get one's Ph.D. by 20."
Two years later, at an age when most college boys are guzzling beer or studying sorority girls, Wolfram became the youngest winner ever of a $128,000 MacArthur Fellowship, also known as "genius awards." There wasn't a field Wolfram couldn't master—high-energy physics, mathematics, computing, artificial intelligence, even cosmology. "I discovered the inflationary universe," Wolfram said. "But I didn't understand its full significance." Perhaps that was what Wolfram feared most about science: mastering the specifics but in the end never capturing the whole.
Wolfram moved on to the Institute for Advanced Study at Princeton University in New Jersey, which had before known the likes of Albert Einstein and J. Robert Oppenheimer. Next, Wolfram landed at the University of Illinois at Urbana-Champaign and was given his own kingdom, the Center for Complex Systems Research. And why not? In his 20s, Wolfram had been a pioneer in the erudite field of cellular automata. Wolfram was figuring out how the processes of other brains, computers and even flowing fluids could be broken down into the interactions of simple, identical parts.
Despite his growing achievements and worldwide reputation, by his mid-20s Wolfram was beginning to tire of depending upon the money and approval of others to carry out his research. The feeling was often mutual. According to John Gage, science director for Sun Microsystems Inc. of Mountain View, Calif., which has been supplying Wolfram with computers for years, the young Briton is "totally impatient, antagonizing to everybody, and very smart to boot, so he can be irritatingly correct."
Wolfram's problem was figuring out how to avoid having to answer to somebody, anybody. And, as usual, he knew the correct solution—software, specifically mathematics software.
Math had always been a large part of Wolfram's work. At Caltech, he had tried to concoct a program that would do the math for his research in theoretical physics. Based on that early work, Wolfram started writing Mathematica in late 1986. The next year he founded Wolfram Research, Inc. in Champaign and began hiring the seven scientists and mathematicians who helped him write the program.
Writing software was nothing like Wolfram's work as a scientist. "Science is analytical, whereas software is very synthetic," he said. "You start with nothing."
Not surprisingly, the scientist took a unique approach to designing Mathematica. First, he tried to map out how the manual would read. If you can't say it in the manual, you won't be able to do it in the program," Wolfram said. As he wrote, he churned out code, often at the phenomenal pace of a thousand lines a day, working until his usual 6 a.m.
Seven hundred and sixty-eight pages later, Wolfram had written "Mathematica," a book published by Addison-Wesley to explain the concepts of his program. At the same time, he'd been chipping away at a good software design that was "relatively easy to implement." Considering its power, Mathematica does seem relatively easy to use: high school students already are using it for algebraic equations.
But Mathematica is more than an educational tool. Math wizards have always experimented with formulas and equations in their minds or on paper. Now mere mortal engineers and scientists can use Mathematica as a calculator to do symbolic and algebraic operations. Graphics can be created with Mathematica to illustrate solutions or processes like acceleration. And the nearly 1 million college students who take calculus each year are expected to benefit from being able to see the answers to their problems.
Wolfram senses this untapped potential and, like the child prodigy and problem child he has been called, he often seems impatient for dessert. At the recent unveiling of Mathematica in Santa Clara, Calif., an umbrella-covered cart was parked in the luncheon room, brimming with the chocolate ice cream bars Wolfram loves. But before he could succumb to his chocolate addiction, he had to attend to business.
Pacing the stage like a nervous boxer, he announced to the sea of suits and flashing lights, in quick, jabbing proclamations, that his software will be available on supercomputers, IBM's PS/2 line, the Mac, and Sun Microsystems and Silicon Graphics workstations. Then, Wolfram sat still for a moment in front of the Mac II, cocked his head to the audience of nearly 200, thrust out his arm, and declared that the Mac version was ahead of the competition. Not the most discreet, diplomatic words, considering that folks from IBM, Sun and NeXT—including Steve Jobs himself—sat only a few feet away.
The next morning, clad in the same sandals and peach socks he had worn the previous afternoon, Wolfram sat in the corner of a Palo Alto, Calif., computer store and explained why his high-level software works better on the Mac than the other computers. "In UNIX, there's not a standard mechanism for editing text," he said. "That's why we haven't been able to design a good interface for the Sun."
So, a listener surmised, Apple helped make Mathematica's Mac interface possible. "Hell no!" Wolfram replied. "We talked once. They suggested that menu items [that are] not active should be gray." It was Theodore Gray, whom Wolfram plucked out of graduate school at the University of California, Berkeley, who spent more than a year making Mathematica "feel right" on the Mac.
Apple's response to Wolfram's charges came from spokeswoman Brooke Cohan. "We worked with them the same way we work with all developers," she said, adding that Wolfram received help from Apple's higher education, technical markets and evangelist groups. "We have hundreds of developers and we try to treat them all fairly." Help or not, Wolfram was befuddled by the standard approach to Mac interface design. "I find it amazing. Everything [in the interface] is argued on precedents," he said.
Apple may have turned down Wolfram's bundling proposal, but he is counting on the company to "help us in reaching Apple dealers." Early indications look good. Next month, Mathematica will be demonstrated in Apple's booth at Siggraph 88 in Atlanta, the annual showcase for computer graphics.
Besides writing the core software of Mathematica, Wolfram uncorked his uncanny knack for finding bugs. "I can often look for two minutes at code, ask a couple of questions and find the bug," said Wolfram, who compared it to unraveling a puzzle. "Sometimes it's a weird thing like asking, 'Why is that array of such and such a length?'"
Despite Wolfram's search for perfection, he said there is plenty of room for special features of Mathematica to be accommodated by third-party developers. "Mathematica provides PostScript output," he said. "I'd like it if you could read it into an illustrator program."
Meanwhile, third-party Mac developers are only months away from releasing what Wolfram called "alternate metaphors for Mathematica." Wolfram said Mathematica will become a more generic tool when developers devise custom methods of using Mathematica, a sort of Lotus 1-2-3 for math. He knows that for Mathematica to grow as a tool, the first step is acceptance by a large audience. Traditionally, math software has found few users. But then, as Stephen Wolfram knows, math software hasn't been designed or sold before by Wolfram.
In just a couple of years, Wolfram has made the jump from scientist to programmer to marketer. He knows the last role will be the toughest. Only a year after opening shop, he has 25 employees. And so far, only Steve Jobs, with his yet-to-be released NeXT machine, has opted to bundle Mathematica.
If all goes well, Wolfram may be taking up his pen again soon. He has completed one-third of a book on the science of randomness that he said is "such fundamental stuff everybody will be able to understand it." Next he wants to design a machine in software and hardware that truly "learns from examples." And he also has ideas about leaping beyond today's "kludgy supercomputer architectures" to "intellectually interesting" designs.
But before these and other ideas can take shape, Wolfram believes he has to make millions with math. And the productivity factor of that equation is that engineers, scientists and students now can use computers, including the Mac, to leap beyond generic applications and generic results.
Wolfram said he is finding the software business to be as satisfying as anything he has done. He may not be uncovering new universes, but some feel his software may be opening new worlds for millions of people. Even for a genius, that has to be gratifying. And maybe, just maybe, Wolfram will find it interesting enough to spin us a few more worlds before he moves on to the next problem.