There is a room in La Jolla office building that is so banal yet so significant that it merits a pilgrimage. The building is at the bottom of the hill crowned by the VA Hospital, next to the Mobil station and across the street from El Torito. Unadorned, the room in question contains little more than a table piled with computer gear. But within those metal casings is a version of the Encyclopaedia Britannica that's not only more accurate than the 1995 version of the print set but already contains far more information — well over 1000 more articles — than any printed version of Britannica will ever hold.
And this pool of information will continue to grow. Every second, bits of it are zapped out to users in a manner never attempted by any encyclopedia company: over the global network of computer networks, the Internet. Most of the work required to launch the 226-year-old institution into the electronic age is being done in San Diego County. The transition promises to cost so much that Britannica has admitted it needs an infusion of cash or the company will have to be sold.
The financial announcement issued from the Britannica world headquarters in Chicago, a setting that contrasts sharply with the La Jolla outpost. The headquarters building, a square-shouldered gray tower overlooking Michigan Avenue, isn’t as old as the company’s main product; few existing structures in North America are. But it has some of the gilt-edged bulk that the world has come to associate with the set of books.
The current editor of the Encyclopaedia Britannica, Robert McHenry, says that sometime in the early 1980s he and his co-workers began fielding an increasing number of queries about the precise number of words that make up those books. “It took a while,” McHenry noted in a recent in-house memo, “and noticing that the calls were generally from computer-related firms, before it was realized that Britannica was being used as an informal unit of measure. It has apparently become common to express memory capacity or, more often, data transmission rates in terms of so many Encyclopaedia Britannicas or EBs per second.
If it helps computer jockeys to think of transmitted information in terms of EBs per second, the 44-million-word description doesn't convey how much there is within the books. Consider that one 32-volume set of the printed Encyclopaedia Britannica includes about the same number of words as 140 average textbooks. A single article (the '83 Macropoedia entry for "United States of America") contains 300,000 words — the equivalent of of four or five trade books.
Wlthin the headquarters building also repose reminders of Britannica’s cultural weight. One is on the third floor, in the company library. There, five black metal bookshelves hold a set of every edition ever printed. The three volumes on the top shelf, in the upper left-hand corner, are replicas of the original First Edition, produced between 1768 and 1771. The company displays an authentic First Edition in a glass case up on the ninth floor, but it’s too rare and expensive to be used as a casual research tool. Terry Passaro, who’s been EB’s head librarian since 1972, says that throughout her tenure the company has had a standing request at Blackwell’s (the famous Oxford bookstore) and Sotheby’s (the auction house) to be informed about any real first editions that surface anywhere in the world. “And I have never been called,” Passaro says. “Wherever they are, they’re set. They just don’t move.”
Even the replica delights Passaro, however, who points out that the first fat volume contains entries beginning with A and B. The second volume holds Cthrough L. “And then they wanted to get it done real fast,” the librarian notes, so M through Z are crammed into the third and final book.
In that last volume, Passaro looks up “Woman” and finds only “The female of man.” The librarian laughs and says, “That’s about as much as they were going to give us!” About “California,” all the Scottish authors had to offer was, “A large country of the West Indies, lying between 116 and 138 west longitude and between 23 degrees and 46 north latitude. It is uncertain whether it be a peninsula or an island.”
By 1778 Britannica's report on California, along with much else, had expanded. (“The Californians are well-made and very strong,” the new essay informed readers. “They are extremely pusillanimous, inconstant, stupid, and even insensible....”) Ten volumes of the much more ambitious Second Edition appeared between 1777 and 1784, beginning a general pattern of expansion that was to continue for 200 years. Some of these efforts still stand out. The Ninth Edition, for instance, published between 1875 and 1889, draws Passaro’s fond gaze. “In this you get a lot more description of technology because we’re in the industrial age,” she notes. The Ninth Edition also incorporates photography for the first time, and it was one of the most heavily pirated editions because the relatively new photographic process made copying so much easier, but copyright laws had not yet been enacted.
“Now this one,” Passaro says in front of a battered green edition, “we use very heavily. It’s one of our most famous editions — the 11th, published in 1910 and 1911. What was really nice about it is that the articles are extremely well written, especially in the humanities.” (The current Encyclopaedia Britannica's own article about itself concurs, stating that “the rich leisurely prose of the 11th Edition marked the pinnacle of literary style in the Britannica.”)
Although the 11th Edition was printed and largely written in England, ownership of Britannica by 1901 had passed into American hands. The mail-order giant Sears, Roebuck and Company bought it in 1920, and throughout the following decade three more editions appeared, the 14th being very substantially revised. Still, the accelerating growth in the world’s stock of knowledge was bringing the encyclopedia company to a crisis point, and in 1932 it announced that it would no longer allow several years to elapse between updates. Instead it would revise the Britannica annually.
This has occurred ever since, with two exceptions worth noting. In 1974, instead of releasing yet another printing of the 14th Edition, Britannica, with enormous fanfare, announced the creation of a 15th Edition. More than 4000 contributors from all over the globe rewrote all the content (with San Diego journalist Neil Morgan this time taking a crack at capturing the essence of “California"). Philosopher Mortimer Adler, who led the huge editorial undertaking, also devised a radical new structure for the self-proclaimed summary of all human learning. Instead of containing 24 volumes covering topics from A to Z, the 15th Edition consisted of 10 volumes of shorter “ready reference” articles (known as the Micropoedia), 19 volumes of lengthy “knowledge in depth” (the Macropoedia), and a 1-volume “Outline of Knowledge” known as the Propoedia.
One notable omission was that of an index, probably the most disliked aspect of the controversial new work. In 1985, therefore, Britannica again reorganized the books somewhat. But this time, for marketing reasons, the company called the new effort merely the 12th printing of the 15th Edition (rather than the 6th Edition). “Refiling the copyright and changing all the sales material can run into millions of dollars,” one executive explains, “ If we can’t really make hay out of such a change, it’s not worth doing.”
The most recent version of the Encyclopaedia Britannica, the one that bears the 1995 date on its title page, thus is still known as the 15th Edition. Its content, however, resembles that of the much-ballyhooed 1974 product far less than a number of earlier separate editions resembled each other. “People know that it’s very expensive to create an encyclopedia,” says EB’s president Joseph Esposito. But he says what they tend to overlook is the work and expense involved in maintaining an existing one. He says publishers of high-quality reference material typically figure they have to spend anywhere from 5 to 12 percent of their original development costs every year to update the material or else their products will quickly seem so outmoded as to be worthless.
To get a better sense of just how many curves the course of events throws at the hapless encyclopedia publisher, it’s helpful to consult the American Library Association’s “Reference Books Bulletin.” The association, which has regularly reviewed American encyclopedias since 1930, considers many elements, but one of the most important is what it terms “currency.” Examining the 1994 Collier’s Encyclopedia, for example, the ALA judged that Collier's had “done a good job of keeping up with recent changes. For example, there is mention of the floods in Illinois during the summer of 1993; ousted Haitian President Aristide’s address to the United Nations in October 1993; the appointment of Ruth Bader Ginsberg to the Supreme Court; the signing of the peace agreement between Israel and the PLO in September 1993; the election of Jean Chr£etien as Canadian prime minister in November; the flight of the space shuttle Discovery in September; and the attack on the Russian congress building by troops loyal to Yeltsin in October.” However, the Collier’s “article on the telephone makes no mention of cellular phones; the latest advance discussed is the introduction of direct-distance calling,” the bulletin notes with disapproval. “There is no reference to air bags in the list of safety features in ‘Automobile’.... ‘Civil Defense’ still illustrates how to build and stock a basic fallout shelter."
The 1994 edition of the Encyclopedia Americana still listed 20-year-old population figures for Spain and referred to punch cards in its article about computers, the bulletin points out. Funk & Wagnalls New Encyclopedia failed to mention the Internet in its article on “Telecommunications.” Britannica, while judged by the ALA to be “the most scholarly of any general encyclopedia,” nonetheless hadn’t gotten around to including any discussion of AIDS in its 1994 “Sex and Sexuality” article. And Britannica's » take on “Telecommunications Systems” still “refers to ‘push-button dialing now being " introduced.’ ”
Beyond the vast number of things in the world that change every year, other considerations further complicate the encyclopedist’s task. A crucial one has to do with the nature of printed Z material. Indiscriminate changes and additions ; to the content of a printed document can require that almost every page be redesigned and reprinted, as type is shifted from one sheet onto succeeding ones. Instead of incurring that enormous expense every year, encyclopedia publishers select only certain sections for modification while resolving that the majority of pages won’t be touched. The new material can then be “interleaved” with the unchanged portions. But which pages should be revised? “That’s one of the editors’ challenges,” one Britannica executive explains.
“If we’ve got 4000 or 5000 pages budgeted to be revised in a given year [out of about 32,000], which ones are going to get the works?” World events can also disrupt this planning process. EB officials say that when Germany was reunited and the Berlin Wall fell (in November of 1989), the company management decided to amend the necessary articles to reflect those events in the 1990 printing, even though the cost would far exceed what had been budgeted for that year. “Not only were the changes made about Germany in the article on Germany,” says a spokesman, “but throughout the set any references to an existing country of East Germany or West Germany — those all had to be changed. And the maps. Altogether almost 5000 pages were affected.”
Even much more routine Britannica updates trigger some surprising consequences. “One of the nasty little secrets of the print encyclopedia,” confides editor-in-chief McHenry, “is that if, as we do, you decide at the outset that there are certain categories...that you will cover exhaustively — say, presidents of the United States or Nobel Prize winners — then you can anticipate that every year there’s going to be an influx of new articles (about the latest batch of them]. However, the size of the set does not change.”
McHenry explains that in order to shoehorn a new article into the pages, some other article must be cut. “It typically is another article of roughly the same size, from that same general vicinity. Not necessarily on that page, but from one nearby in order to minimize the number of pages that are in work. Over the years, as more and more Nobel Prize winners and presidents and British prime ministers and all that sort of thing pop up, you find yourself sitting on a pile of articles that used to be in the set but have been taken out, not because you wanted to or they’re somehow less useful, but simply because something had to go.”
This constraint, once again, springs from the nature of print. And McHenry points out that traditionally the printed books and the very concept of the encyclopedia were “coextensive — one and the same, identical, philosophically equivalent in every sense.” He says that six or eight years ago, when it first began to dawn on him that this might not always be the case, that the Encyclopaedia Britannica might one day shed its print straitjacket and take life in some far more fluid electronic form, the prospect struck him as being "just heaven.. .Valhalla.”
Today Britannica has two electronic incarnations: one a CD-ROM and the other the on-line version accessible over the Internet. But both have just become available within the past year, and the on-line Britannica isn’t yet being sold to home users.
Did it take too long to develop these products? Some observers have posed the question in an apocalyptic context. In February of 1994, Forbes magazine went so far as to call Britannica “CD-ROM’s First Victim," bluntly declaring that EB’s tardiness in embracing the new computer technology had already “wrecked” the 200-year-old company. The central point of the Forbes article was indisputable; Britannica had thus far missed the boat that was carrying several other encyclopedia companies to fame and at least the prospect of fortune. But the article also contained a number of serious factual errors, and it gave the erroneous impression that Britannica by 1994 had turned its back on the electronic realm.
What really happened was a lot more complicated, as Harold Kester tells the story. Kester is the central star in Britannica’s La Jolla constellation, director of the “Advanced Technology Group" there. At 48 he moves quickly, talks quickly. He wears his iron-gray hair plastered straight back, and this shows off a broad, high forehead that descends to a pair of round metal spectacles. Below them, an unlined, impish face. One recent morning, Kester dressed in blue jeans and a dark jacket over a white T-shirt emblazoned with Chiat/Day’s ominous dictate, “Innovate or Die (and death is not an option).” He would have looked at home at any Hollywood power lunch.
He says he lived in Los Angeles for a while after getting a math degree from Cal State at Long Beach and cofounding a circuit design software company. But when he was 13 years old, he had visited San Diego, and one day, while driving down to Torrey Pines Beach from Del Mar, he had concluded, “This is paradise.” When his company was bought in 1972 by a larger Rancho Bernardo software firm, he rejoiced. “They told us on a Friday that we’d been purchased, and literally by Sunday I had bought a home in San Diego.”
The work as well as the scenery here satisfied him for a while. Eventually TRW acquired the Rancho Bernardo software firm, and it later joined forces with the Japanese electronics giant Fujitsu. But Kester began to yearn to do something more entrepreneurial, and by 1984 he had discovered a product idea that excited him.
Conceived of by a professional associate named Bob Clarke, the notion was to develop and sell to bookstores a device called the Fiction Advisor. “It would ask you some questions,” Kester says today with a wistful affection.
Based on your touch-screen responses, “it would recommend what books to read. It was really quite good. Anyway we started a company based on that.”
They named their venture the Del Mar Group (since both Kester and Clarke then lived in Del Mar), and they raised half a million dollars (their own money and friends’) to fund its startup. But Kester says after he had quit his job, a key investor pulled out. “So basically I got behind the power curve in terms of financing.... And I never really caught up.”
It also soon became clear that the Fiction Advisor had some crippling limitations. “We were so naive,” Kester moans today.
“It really was trial by fire.”
Although people told him and Clarke that they liked the Fiction Advisor, the problem it solved “was but a small percentage of the total problems of bookstores,” Kester says. If you walked into the bookstore wanting advice on which novel to read, the Fiction Advisor could offer some nifty suggestions. But if you walked in looking for a book about anorexia nervosa, “This product wasn’t for you.
“What we needed to give them was a certain kind of information-retrieval system to match up with the inventory that was there in the store,” says Kester. For that, “We needed content.” So they managed to strike a deal with the publisher of the Books in Print reference work to put its database on a CD-ROM. “In 1985 these were just coming on the scene,” Kester recalls. “Before that there was no way to store a lot of information other than being networked, which was too costly.
“We produced the fifth CD-ROM that was published in the U.S., as far as we know,” Kester boasts. “It cost us $10,000 to get it mastered in Japan.” Customs officials almost wouldn’t let them bring the (then) mysterious object into the country. Kester says this product was “a sort of prototype for what is now the Books in Print on disc. If you go into a bookstore today and you see a computer there, and there’s a CD-ROM in it, it probably has Books in Print. Anyway, Del Mar Group did the first ones.”
In the fall of 1985, they added some digitized pictures of book covers to their product, making it one of the first “multimedia” discs ever produced. Kester mentions the accomplishment while conveying some disgust for the hype that has come to surround that word. He makes it sound as if the far more exciting development was what he and his group learned when they started watching people use their creation. “We discovered right away that the person who approaches a database for the general consumer is not an information-retrieval expert. They probably don’t know Boolean logic. They often misspell words.
“We actually went out and studied what people would ask for at the information desk,” Kester continues. Today he has a collection of some of the more bizarre associations. One customer, looking for a work he’d heard about on public radio, asked for How to Press Rats when, as the store clerk eventually divined, what he wanted was Oedipus Rex. “Now this person obviously was not familiar with Latin, and when you approach new languages, all you hear is noise,” Kester explains. “In that context, the error is not that weird. Another example: customer asked for Dark Angel but actually wanted Black Devil.” Still another wanted The Amityville Horror and asked for The Amityville Whore.
Over time, the Del Mar Group got caught up in the broader challenge suggested by the book searchers’ experience. “The problem that we tried to solve in the book business got us into the generic problem of text retrieval — basically, how to find what you’re looking for in very large textual databases.” As Kester continued hustling investment money and the company took on other consulting work — “anything and everything to get money to keep the place alive” — they also began creating a new product. Dubbed Smartrieve, it would do more than simply retrieve text that matched the exact words typed in by the user. Kester’s team instead vowed to build into the search software some intelligence, so that it could better respond to the human users’ many creative and unexpected ways of looking for information.
Kester says by 1988 he had secured venture capital, and the product was ready to be launched. Then Kester’s son committed suicide one week before his 18th birthday. “Within 13 days, my wife and I were separated, and I was essentially bankrupt," Kester narrates. “What happened was that as soon as Kevin committed suicide, the venture capitalists said, ‘This guy is going to be unstable,’ and so they wouldn’t put their money in.”
Kester discloses these facts by way of explaining how it was that he was so dazed and befuddled that he almost failed to make his fateful connection with Britannica. He says a month after the suicide, he was in Seattle presenting Smartrieve at Microsoft’s CD-ROM trade show when another San Diegan named Greg Bestick strolled up to his booth. Bestick worked for Education Systems (now jostens Learning), a Mira Mesa-based creator of computerized curricula being used in elementary and junior high schools. The company’s managers wanted to link a reference work to their networks of classroom terminals, and they had thought immediately of the Encyclopaedia Britannica. A collaboration between Britannica and Education Systems was being negotiated when Bestick flew to Seattle.
“He knew the kids weren’t going to be information retrieval experts," Kester explains. “So he wanted to have the most consumer-friendly search-and-retrieval engine that he could find to search the text. When he saw Smartrieve, he felt the Del Mar Group had this.” Yet Kester was in such an emotional fog, he says, that when he returned to San Diego he didn’t remember anything he and Bestick had talked about.
Apprised of Kester’s personal problems, Bestick persisted, contracting with the Del Mar Group to design how Smartrieve would work with an encyclopedia — specifically, with the Compton’s Encyclopedia, rather than the Encyclopaedia Britannica. This choice made sense for a number of reasons. Although Compton’s, back in the 1950s, was a major encyclopedia publisher, it had fallen on hard times, and Britannica had bought it “decades ago," according to EB president Esposito. By the late 1980s, “The (Compton’s] business was really moribund,” he says, “it existed at that time solely to be used as a premium to help sell the print Britannica.’’ If potential buyers fretted that the scholarly Britannica might be over the heads of their offspring, the salesman countered that the youngsters could start with Compton’s, then “grow into” the bigger set.
Putting Compton’s on a compact disc thus in no way threatened Britannica’s print sales. And Compton’s was geared to schoolchildren — the audience served by Education Systems. Moreover, it was a much smaller corpus — some 9 million words versus EB's vaunted 44 million. Getting it to fit on a compact disc posed no serious technical challenges, even as Education Systems added thousands of electronic appetizers to the informational banquet — such things as a snippet of Bach’s Brandenberg Concertos, a clip of Richard Nixon denying his crookedness, animation showing how a skeleton moves. “The outcome was that in April of 1988, Education Systems, with EB, announced that first product,” says Kester. Although Grolier’s had produced a text-only version of its print encyclopedia back in 1985, the Compton’s disc represented the world’s first multimedia encyclopedia, and it commanded widespread attention.
Why didn’t EB follow up by rushing out a CD-ROM version of the main encyclopedia? Kester contends that the company, in fact, took one of the first steps toward doing so within a few months of the Compton’s CD debut by commissioning the Del Mar Group to develop an electronic index to the main set. “You could ask it a question, and it would tell you where [to look for an answer] in the print set,” he explains. Kester also argues that the commission revealed Britannica’s foresighted recognition “that electronic publishing was in its future. Nobody knew in the late 1980s how to sell it or what the business model would be.... But they were absolutely making the commitment to understand this medium.”
At the same time, both technical and cultural forces were acting to keep that future at bay. The technical problem stemmed from Britannica’s enormous size. As roomy as a compact disc is (holding almost two-thirds of a billion bytes of information), it’s really not quite big enough for the Big Mama of the reference world. While her 44 million words could be encrypted down to about 130 million bytes, the index to those words plus other electronic overhead took up about 810 million. And that was with none of the frills that so charmed consumers of the CD Compton’s and the competitors (like Microsoft’s Encarta) that followed it — no byte-greedy music or pictures or video.
Of course, as every overweight matron knows, girth can always be squeezed and trimmed if the spirit is willing. But for some time, countervailing spirits within the Britannica establishment were horrified by the vision of their majestic product, for which buyers routinely paid $1500, being delivered on a little plastic and aluminum platter commonly perceived to be worth $15 to $20. Reported Forbes in its February 1994 article, “Several former executives who recently left the company say Britannica didn’t want to risk offending its powerful sales force.... On CD-ROM, Britannica could have been priced much lower than the paper version. But the lower the price, the lower the selling commissions.” EB president Esposito acknowledges that “there’s no doubt that up until a certain period of time, a lot of decisions were being made to defer to the direct-sales force. That was true.”
What Esposito adds, however, is that by the end of 1993 — two months before the Forbes article reached print — dramatic changes had already rocked Britannica. “The real story that the (Forbes] article missed was that the preceding September, Britannica’s CEO of the last 20 years retired. Peter Norton became our new CEO. And within three months, 50 officers and directors of Britannica were out the door.” On January 1, 1994, the company was reorganized, with Esposito taking command of all the North American operations, and since then, “There is no aspect of our operation that has not been changed somewhat,” he declares.
“We’ve completely reorganized the sales organization. We’ve changed the whole marketing model.... There’s no cold calling. We used to advertise on television, and people then called up, and we’d take those leads and try to sell them (encyclopedias). But you could never make money on such a shotgun approach to the marketplace,” the company president states. "People who don’t have a certain degree of affluence, people who are not interested in education, people who are not themselves college-educated, who don’t have aspirations for their children — these probably are not strong candidates for Britannica. But the television advertising picked up everybody.
“So we cut that way back,” Esposito continues. “We do much more with targeted direct mailings now.” Also, “The bulk of direct selling does not take place in people’s homes. Half of our business is now at counters and at trade shows.” He explains, “Let’s suppose you’ve got 10,000 podiatrists meeting in town. In the exhibit hall, there will be a Britannica booth staffed by a Britannica sales rep who’ll do a presentation there. And we’ll sell you the print set, the CD-ROM, the ‘Great Books....’ ”
In the future, that sales rep will be offering even more products bearing the word Britannica, promises Esposito, who believes that it is now “the most underexploited brand name in America.” The company has become promiscuous, Esposito likes to tell reporters. “If we found a way to market (the encyclopedia] on audiotape, we would do it. If we found a way to market Britannica by stenciling it to the windows of airplanes so that every time you looked out the window you could read an article from the Britannica, we would do that.”
That’s not yet possible, but by the end of 1993 Esposito did give the nod to Kester’s group to put the Encyclopaedia Britannica on a single compact disc. (A cumbersome two-disc version had been introduced in the fall of 1993, aimed primarily at publishers and other corporate users who had an urgent need to check facts reliably.) To create the streamlined consumer CD-ROM, the La Jolla team eliminated some obscure, rarely used searching capacity from the earlier product and did some other minor tinkering. The resulting product began shipping in July of last year. Esposito says that now, if you talk to the sales force, the CD-ROM “is their favorite thing. They love it.... It’s a growing percentage of our sales every month, and most of the customers are buying CD-ROM in combination with print. They’re buying packages” — paying a higher price to get the compact disc in addition to the books.
Why would anyone want both? That’s not hard to understand when you see them side by side. Looking something up in the print volume is a sensual experience. The fine-grained paper slides like satin underneath your fingertips. The print, though small, soothes the eye with its anthracite crispness. You can curl up with a volume in an armchair, a glass of wine at hand, your children cuddled around you, and nibble on the information like a Roman emperor enjoying honeyed figs.
But suppose what you want from the encyclopedia is an answer to a question such as, “Why does the moon loom larger on the horizon than it does high in the sky?” Where to look in those hefty print volumes is not at all obvious. Typing the question into the Britannica on CD-ROM, however, yields a list of articles that might be relevant to the answer. Then you’re only a mouse click away from the scholarly explanation embedded in the essay on “Human Perception.”
“I find I ask different questions of each medium, and I use it for different reasons,” says Esposito. “Now, let me be very clear about something. We’re completely agnostic as to what people use. If someone wants to read a 200-page article on China from their screen, go right ahead. Not our problem. If somebody else wants to answer a question like, ‘Why is the sky blue?’ from the print set, have a good time. It might take them six months to find it. But that is not our problem. Our problem is to make it available in whatever form people want it.” And well before the Britannica CD hit the street, Esposito and other key figures within the company had become convinced that people will soon want to get encyclopedic information in yet another manner — by having their personal computer connect with a remote “information server.”
Harold Kester says Bob Clarke (“our technical visionary”) had come to believe this as early as 1991. “Bob reads everything every day,” Kester declares with some pride. “He has a major in philosophy and he’s a literary person, but he’s been in the (computer] business for about 35 years, and he really has the ability to see where technology and the markets are going.” Clarke himself recalls that the growing power of personal computers inspired his vision. Armed with increasingly smart terminals, future encyclopedia customers could tap into an encyclopedic database of unlimited size, one that could be updated constantly and linked to a staggering array of other information resources.
By early ’92, the La Jolla group had marshaled their arguments about the future of encyclopedias and shipped them off to Chicago in a position paper that Kester says received “more than casual interest.” Other events in Chicago also nudged Britannica’s managers into thinking about delivering their central product over a computer network. The University of Chicago, the sole beneficiary of profits earned by the foundation that owns EB, expressed interest in offering the encyclopedia over its campus-wide computer system. “We started to explore that, and we discovered that it was going to cost us a couple of hundred thousand dollars to get it running,” says Esposito. “But then if we wanted to sell it to UCSD, it would also cost us a couple of hundred thousand dollars to get it set up there, and so on. We would have this big cost every time we wanted to do it.... We found out that, at best, we could get a payback after 10 or 12 years.”
Dismayed by that prospect, Esposito’s recollection is that he turned to Kester, who was by then a Britannica employee. (In 1990, exhausted by their funding struggles, Kester and Clarke had sold the Del Mar Group to Britannica, which had merged it with another software development group and dubbed the resulting entity Compton’s New Media.) Esposito claims he phoned Kester and asked him to investigate whether a network-based product could be developed once, then resold over and over again. “Harold now disappears, and the project is in limbo,” Esposito recalls. “But some time later, I get a phone call from Harold, really agitated. He says, ‘Did you ever hear of the Internet?’ ” Esposito says he answered, “Who? What?”
Such ignorance was not at all unusual. The Internet had come into being in 1969, when certain government and private agencies contrived a way to connect their in-house computer networks with each other via phone lines. Nonetheless, even a few years ago it remained the almost exclusive province of people armed with great stocks of both computer expertise and patience. To do anything on it, the majority of users had to type commands in the dense, daunting language of the Unix operating system. And what streamed across their screens when they connected was bare, unadorned text.
By late 1992, this situation was just beginning to change. A Geneva-based networking expert in 1990 had conceived of the World Wide Web, a model for processing documents in such a way as to simplify the task of viewing, linking, and electronically publishing those documents. Software tools for making the model a reality were just beginning to proliferate. These developments were sufficient to inspire certain key people within Kester’s group with the vision that this was the way to link paying customers with Britannica.
Without question, the alternative — for Britannica to establish its own proprietary computer network — would have been expensive, complex, and a serious diversion from the company’s main business, imagine a phone sex business having to first set up its own phone company over which to deliver the goods. It is far easier to offer what you’ve got over some existing delivery system. But two and a half years ago, could the Internet be considered such a system? It was evolving on an almost daily basis. No one was doing business over it; rather, a self-righteous anti-commercialism pervaded the user community.
Small wonder, then, that Chicago greeted the La Jollans’ brainstorm with something less than wild enthusiasm. Undeterred, Kester’s group proceeded to develop the idea, and, Kester says, by May of 1993, Britannica’s management had come to share the California group’s excitement. In October, another key component fell into place as the Mosaic “Web-Browsing” software became available to users of Windows and Macintosh machines. Armed with such software, personal computer users could suddenly view World Wide Web documents in their full graphic glory — not as naked character streams, but garbed in headlines and various fonts and ornamented with complex images. Rather than typing mind-numbing computerese, they could connect to other places on the Internet by placing the cursor on an icon or a highlighted word (a so-called hot link) and clicking the mouse. Travelers on the information highway could trade their ox cart for something at least resembling an automobile.
Kester says the task of securing a good “search engine” for the on-line product also turned out to be painless. He says Smartrieve wasn’t an option. Britannica by the fall of 1993 had sold off Compton’s New Media to the Tribune Publishing Company, and for its $57 million, the Tribune company got to keep the search software that the Del Mar Group had developed. But far more significant was the fact that Smartrieve wasn’t designed for a client-server architecture. Instead the La Jolla group discovered they could use the database searching tool developed by the Wide Area Information Servers (WAIS) company in Menlo Park, a tool that, given some refinement, would do what they needed.
So it was that — less than a year after committing to the idea — Britannica had a version of itself on the Internet, and in February of 1994 certain people on the UCSD campus began to have access to it. The La lolla campus was a logical choice for a “beta test” site. For one thing, Kester had forged a relationship with UCSD computer science professor Rik Belew and several of Belew’s graduate students (who work for Britannica today as consultants). Furthermore, “The UCSD library has a history of doing this kind of work,” says associate university librarian Bruce Miller. “We like to push the boundaries.”
Miller says that before the on-line encyclopedia reached the campus, he and Belew had discovered an unexpected affinity between Britannica and the UCSD library. He says Belew’s students compiled a random sample of more than 1500 of the bibliographic citations that appear at the end of the EB articles. Then the library staff compared the books cited against those in the campus library catalog. “Boy, were we surprised!” Miller exclaims. The research project showed that 69.3 percent of the books the Britannica authors recommended for further reading were in the collection housed on the La Jolla.campus. Another 8.5 percent were almost the same, differing only by edition.
“This was totally nonintuitive,” says Miller, who explains that university research libraries “don’t regard encyclopedias very highly.” They’re seen as superficial summaries, rather than serious scholarship. But the Encyclopaedia Britannica defies the stereotype. Miller asserts. “It really offers substantial background on each of the things it covers. Kids usually need an adult mediator to use it effectively.”
So Miller became enthusiastic about testing the on-line incarnation, though the campus began that test with great restraint. “When we brought up Britannica Online, we brought it up in a way that made it almost assured that we would have the fewest number of users connected to it.” Only members of the UCSD community who had personal computers equipped with World Wide Web browsers could tap into the service. “And in the spring of 1994, getting a Web browser was difficult, and installing it was next to impossible,” the librarian says. Little effort was made to let people know that Britannica was available on-line. “And then came the summer, when typically no one is here.”
He was thus impressed when during that time period, Kester’s group analyzed the response to their creation and determined, “We had 2000 uses a month.” Word of mouth alone was attracting the users. “Now that means something to me,” Miller says. He also found the attraction understandable. “In my entire time here at the library, I have never gotten up and used the (printed] Encyclopaedia Britannica in my work.” But once he could consult the Britannica’s contents with just a few clicks on his computer terminal, he began doing so. “For example, I’ve been interested in some issues of international copyright. I had lots of professional literature on it, but it was all very United States-centric, and I just wanted to see what the rest of the world was doing.” Miller says he found in the online Britannica “a very substantial overview article that really got me on the right track. It was almost like reading a small, very focused monograph.”
As segments of UCSD’s wired community explored the prototype, the Britannica staff continued working on key refinements. Although most of the articles had been converted to electronic form as long ago as the late ’60s and early ’70s, some gaping holes remained in the database. “None of the photographs were digitized and none of the illustrations,” says Neil Holman, Britannica’s vice president for product development in Chicago. Moreover, “Many things which might look like they’re just characters on a keyboard are not,” Holman adds. Greek characters, music notation, chemical notation, mathematical symbols — all these and more had been treated as if they were artwork, stripped in where necessary within the text. Holman says at one point Joseph Esposito wondered if “perhaps we could get away with just putting the text up (on-line). I said we couldn’t do it. There’s too much meaning lost if you drop those things that are essential:” Deletion of the formulas might turn many a math or science article into gibberish. “So the editors were assigned to decide what is essential. What do we keep?” said Holman.
The editors also scrutinized Britannica's 23,000 drawings and illustrations and decided that about 3000 of these added indispensable information to the words they accompanied. But these couldn’t simply be scanned and fed into the computer. “We tried that,” says Anne Long, Britannica’s executive director for electronic products, “but we just hated the results. Our black-and-white line art is very finely done. It’s very detailed.” What looks elegant on paper appeared almost indecipherable on a computer screen, so almost all of the 3000 targeted illustrations were re-created for the different environment.
Last September, Britannica began selling the on-line service in the form of a yearly license available to college campuses. Pricing depends upon the number of students. A campus with 2500 pays $2500, one with 10,000 pays $7000, and so on. Pricing for other marketing channels — public libraries, corporations, home users — is now being decided upon, and Esposito says everything will be announced before the end of the year. But curious individuals need not wait until then to see the on-line Britannica in action. If they have full Internet access, they can ask to join Britannica’s “Early Experience Program” (by sending an e-mail to [email protected]). This is a market research program that gives home users at least 90 days of free access in exchange for answers to a number of questions. Yet another way to sample Britannica Online is to go to the reference section of the UCSD library, sit down at one of the Internet terminals, and make one’s way to the Britannica “home page.” A bold image of the earth against the blackness of space pops into view next to a vibrant red banner emblazoned with the words “Britannica Online.” One click takes the user to a search box into which can be typed anything from a single word (for example, “Shakespeare”) to several sentences. In the case of the latter, as with other inquiries expressed in natural language, the “System Britannica” search engine that Kester’s group has developed first eliminates all the so-called stopwords — items such as “the,” “a,” “of,” “more,” and other expressions so common that their inclusion would bog down the search.
Then it checks its index to find where every remaining word appears throughout the entire Britannica corpus and makes a list of its findings, ranking them in order of their likely relevance. Asking “What year did Disney first create Mickey Mouse?” for example, generates a list containing hundreds of Britannica entries — every article containing the word “year,” “Disney,” “first,” “create,” “Mickey,” “mouse,” or any combination of the six. But at the top of the list is an eight-line entry titled “Mickey Mouse” that provides a succinct answer.
The mechanics of using this encyclopedia differ so much from those of using the print set that some of the differences in the content are hard to discern. But they are substantial. Britannica has resuscitated most of the articles squeezed out of the print set when “Bill Clinton" and “Nelson Mandela” and “Kary Mullis” and their like were shoehorned in, according to editor-in-chief McHenry. So have about eleven articles from previous editions that Britannica deems to be classics. “These are all articles that have been superseded in some sense but retain their interest either for their literary quality or their historical value or just the fact that there was this wonderful coincidence of time and subject and author,” McHenry explains.
For the 13th Edition, for example, Albert Einstein contributed an essay on “Space-Time” that was later replaced with information from more modern physics authorities. But now Einstein’s view of the subject can be consulted again, as can H.L. Mencken’s discussion of the American language, Chesterton’s loving portrait of Dickens, Freud’s summary of psychoanalysis. McHenry says that roughly 200 articles fall into this “Classics” category, and current plans call for eventually bringing them all into the electronic pool.
Not only old material is being piped in to augment what’s in the print set. McHenry says a handful of brand-new articles that have not yet found their way into the bound volumes can now be accessed through the electronic gateway. And Britannica has commissioned about 3300 new articles — one about every county in the United States and an additional 300 or so about academic institutions. On-line users also can comb Britannica’s Books of the Year, the annual summaries that update older print sets and also contain more ephemeral material than would ever make it into the main encyclopedia. Want a precis of “grunge”? Although the word appears nowhere in the main Britannica, the 1994 Book of the Year discourses on the fashion and musical style in three separate locations.
Britannica Online also differs from all other encyclopedias, both those in print and on CD, in a revolutionary way. Because its database is part of the Internet, users can be led to other far-flung resources that also make up the information web. Already more than 1000 such links have been incorporated into the on-line encyclopedia. At the end of the “San Diego” Britannica article, for example, you find three “Related Internet Resources” — the official City of San Diego Home Page, San Diego information from the Virtual Tourist, and the home page offered by the San Diego Historical Society. A few clicks of the mouse take the user to viewring options ranging from photos of La Jolla homes for sale to a view of the historical society’s archives room to maps of the Lindbergh Field terminal buildings.
Currently the on-line Britannica is revised every four months. Although that’s four times more often than the printed set is redone. Bob McHenry, the editor-in-chief, sounds like he’s already chafing at the inflexibility. He’d like to be able instantly to respond to “major events that would necessitate some sort of revision,” he says. “Obituaries are an obvious example.... We don’t want to be a newspaper. But there are certain kinds of information where, if you’re not saying what is the case, then you’re saying something that’s factually incorrect.” McHenry adds that the next version of the WAIS search engine may give Britannica the ability to do an “incremental build” — that is, to take out a single article, rewrite it, and replace it without having to reparse the entire database and rebuild the index.
If he can’t do that yet, McHenry nonetheless expresses delight over what the on-line encyclopedia promises to make possible. “The encyclopedia’s job has always been, as we understand it, to inform its users of the current state of learning. But there’s a built-in frustration with trying to reflect the current state of learning when you know there’s always a lapse of time.” He offers this example. “I was just signing a letter to a professor at Yale, asking him to write a new article on Tito, the late president of the late Yugoslavia. Let’s suppose he decides he’ll do it. We’re asking him to get us a manuscript by next March. He’ll start writing that sometime this summer. He’ll finish it in the fall or winter and send it to us. At some point, he will stop looking at the research materials — even before he finishes writing the article. So there’s the end of its currency in terms of the latest word on the secret life of Tito. We’ll get that article next March and it’ll be another year before it gets into print — at which point it’ll be a year and a half away from the journal research. And there it will stay. Until we can get around to it again."
The advent of Britannica Online may change this process in more than one way, McHenry suggests. The obvious one is to shorten the lag between the time a manuscript is finished and when the public has access to it. Beyond that, however, “The same tools that permit us to get the article out on-line may also make it easier for us to keep the author in touch with his own work, to keep his hand, as it were, in the ongoing revision.” Future Britannica editors thus may commission more than the mere writing of articles; they could also have the authors review and update the material over time.
“We’ve been having all these visions, and we don’t know what to do with them,” McHenry admits. “We’re now getting e-mail from users. Starting from zero, it’s been growing at roughly a geometric rate for the last four or five months.” Some point out what they believe are errors in the text; others mention additional points that they believe should be added. “Right now we’re saying, very politely, ‘Thanks very much. We’ll attend to this just as quickly as we possibly can.’ ”
McHenry suggests it will take some time for the Britannica staff to learn not only how best to exploit the potential but also to avoid the pitfalls of the on-line medium. Among the latter, he mentions, “One danger is to say, ‘Well, the article doesn’t need to be cut off at 300 words [because that’s the size of the hole on the page]. Now the article can be as long as we want. In fact, they can all be as long as we want, can’t they?’ Yet the constraints that the print medium forced upon us were a discipline. They made you write a really artful summary of some subject in a brief space. It would be a shame to lose the art just because you could now have 2000 words instead of 300.”
He says another “useful discipline” imposed by print was that very lapse of time between the latest scholarly word and the time an article reached print. Such a lapse “is not always a bad thing,” the editor reflects. “Sometimes it’s exactly the time you need to exercise a little judgment. And another of the dangers of the new medium is to rush — not into print — but into bytes too precipitously, without taking time for events to mature and for [informed] judgments to be made.”
Will there also be more danger of factual errors creeping in if that lag time shortens? For the moment, McHenry says the same standards are being used for editing, fact-checking, and verifying all Britannica articles. Yet he can’t resist musing about how the very nature of an error somehow seems different in the print versus the on-line versions. In the print process, one factual error, one instance of ambiguity or inconsistency “becomes at the end of the publishing process 150,000 or 200,000 instances, each likely to mislead or disappoint any reader who happens upon it,” McHenry quotes from the Britannica stylebook. “That’s a fairly chastening thought — the fact that in the printed encyclopedia, you can’t commit one error. Any error comes in multiples of 100,000.
“How does that translate into on-line?” he continues. “Well, we can pretend it’s one error instead of 100,000 because it’s just that one file sitting there.... In a sense, they don’t seem to mean as much. And there’s also the fact that unlike print...we can Fix the on-line error tomorrow.”
With the disappearance of any restrictions on its size limit, Britannica could launch into a big expansion of its on-line multimedia components, but such an expansion doesn’t appear to be imminent. “We haven’t really put our emphasis on multimedia development for a number of reasons,” Esposito says. “Not the least of which is that we think what we primarily have to offer is our comprehensiveness and authority. And that derives more from the text of Britannica. We have nothing against multimedia, don’t misunderstand me. But if you’re sitting there with a fairly small encyclopedia, and you're trying to find a way to sell it, you add pictures and sound right away. If you’re sitting there with a massive encyclopedia, what you try to do is to make that massive text more useful through electronics.”
The La Jolla group is therefore continuing to beef up the System Britannica searching capacities. “We have just begun to apply the knowledge that’s in the corpora of Britannica to make the search system smarter,” Kester says. Take the problem of recognizing that something is a phrase. While the average person knows that the words “Wichita Falls,” for example, refer to a place, the computer needs to learn that, otherwise, when told to search for “Wichita Falls,” it will come back with a list of every article containing not only “Wichita,” but also “fall” and “falls.” How do you educate your search system to such complexities?
“Well, Britannica owns the Merriam-Webster Third International Unabridged Dictionary,"says Kester. “And because we own them, I can go and incorporate their word list into our phrase dictionary. Or another example. Somebody has gone through the Britannica and read it and created a database of every person, place, and thing. This was a research project that was started about ten years ago with no sight in particular of why they wanted to do it.” Now, however, that project has gained a sudden utility.
The company’s work with electronics has also added a different sort of value to its product, Esposito argues. Say you’re Princeton University, he proposes, and you want your 10,000 students to be able to use the Encyclopaedia Britannica. You can buy the print set for $2000 and store the volumes in the campus library. But how many students will actually use them there? As an alternative, you can pipe the Britannica Online into the room of every single student for $10,000. “Is that an increase in price or is it a decrease?” Esposito asks. He answers his own question by asserting that the on-line product is “probably the most economical way to get a reference database that there is. Because what we’re doing is adding value through distribution.”
Six months into the life of the on-line product, that pitch didn’t appear to be dazzling universities. At the end of March, Esposito disclosed that only ten American campuses (not including UCSD, which remains a beta test site) and one European school had signed up. A month later, however, “several more colleges” had signed on, and Esposito was defending the response rate (in an e-mail response to a query about it) by saying it was “ahead of forecast.” It is “remarkable by any standard,” he further contended, “since college libraries typically have to wait until the following year’s budget to buy a new product (i.e., they have to plan the purchase). We will have more than a million college students using BOL by the end of the year. We are too stately to say we are ecstatic.”
In any case, Britannica’s survival won’t rest upon how the institutional sales of Britannica Online go. “Britannica is a consumer products company,” and “over time, Britannica Online will become the core Britannica encyclopedia,” Esposito has declared. So it is home users who will have to embrace the on-line product. If they don’t, the company isn’t likely to reach its 250th birthday.
At the moment, Britannica can report that consumers at least seem curious. Electronic products director Anne Long says she’s felt “enormous pressure” from people wanting the on-line access. “I come in every morning, and there are between 30 and 50 e-mails and half of them are ‘Can I get this?’ ” she says.
Esposito says the answer to that question will be yes before the end of this year. Some answers to how the on-line consumer service will work have begun to emerge. For example, John Dimm, a senior software engineer in La Jolla, says he and his colleagues dislike the idea of basing the customer charge upon the number of documents downloaded “because that would mean you would feel restricted in what you could look at.” For that reason, Dimm indicates that support within the company has grown for “a subscription service which is by the month or by the year — within a certain kind of limitation so that we know you’re not a robot downloading the whole thing.”
Britannica spokesmen aren’t saying anything about how much a subscription will cost the average Joe, but if the Britannica’s history of pricing its other products is any guide, it won’t be cheap. Prices for the print set currently range from about $1000 to $2500 (depending upon whom you buy it from and what you get along with it). Inexpensive competitors have always been around, “Yet the print set in its sheer magnitude, its august size, subliminally communicates the notions of comprehensiveness and authority,” Esposito says.
When Britannica introduced its consumer CD-ROM last summer, it bore a $995 price tag — compared to less than $ 100 for the likes of Compton’s, Encarta, Grolier’s. Says Esposito, “The real cost of the product has less to do with the medium it’s published on and more to do with the ability to amortize the development costs. If you priced Britannica at what you can buy Encarta for, there are just not enough households in America to pay for the development.... Now I like Encarta. I like Compton’s. I like Grolier’s. These are great products. But what we do does not really compete with them. The only thing Britannica shares with those other products is the word ‘encyclopedia.’ And in case you haven’t noticed, we don’t use that word very much anymore. We don’t own it.” Plus, the word encyclopedia has been trivialized, Esposito contends.
“You can walk into a bookstore and go to the New Age section and find an encyclopedia of spirits and channeling. You can find the encyclopedia of herbal cures. That’s not helping the word any.” Esposito says not long ago he commissioned an informal study of various databases over the last ten years. “It wasn’t really a scientific analysis,” he says, “but it did confirm what we had suspected. The word encyclopedia was being used three times more often now than it was ten years ago.”
Britannica, in contrast, is “reinventing what an encyclopedia is,” the company president suggests. One can argue that Britannica’s sales force has succeeded at selling some variation of that line to consumers for much of its history. But now the company must also sell it to outside investors. In its April 4 announcement, Britannica explained that it was “now in the process of identifying new sources of capital.... We’re confident we will secure the financial resources, which might be in the form of joint venture partners, outside investors, or even a new owner....”
Some of the newspapers that reported on this development did so as if it were both unprecedented and somehow disgraceful. In fact, viewed from the perspective of more than two centuries, it’s more like a tradition. Time after time, Britannica’s publishers have run short of cash, sought more, and rebounded. On many occasions, the whole operation has changed hands — even crossing the Atlantic in one such transfer. If this happens again, won’t it be just another chapter in what’s already a very thick book?
On the other hand, if that computer room in La Jolla does come to attract hundreds of thousands — or millions — of users who collectively pay enough money to support a huge staff of people devoted to keeping the stock of information within the computer up to date, won’t perhaps a whole new book have to be written? Esposito offers this perspective from the midst of the change the company is undergoing. The executive says that not long ago he was contacted by a man who was organizing a trade show about doing business on the Internet. The man wanted Esposito to be his keynote speaker because, he said, he had asked around and had been told that Esposito was the leading expert on that topic in the country. “I said, ‘Let me tell you something. If I’m the leading expert on doing business on the Internet, we are all in very serious trouble!’ ”
Esposito wasn’t joking as he recounted this. “What we have been able to do is to develop a revenue stream already. Not everybody’s been able to do that. But we don’t know what we’re doing! We are making it up as we go along.”
How satisfying is it to use the Britannica Online? That depends a lot on where you’re sitting.
From the seat at my home computer, the time lags involved in getting to BOL and, to some extent, extracting information from it make the experience pretty unappealing. I’m using obsolescent, though fairly common, equipment: a Packard Bell 486SX 25MHz machine with 8 MB of RAM and a fast 28.8 KB modem. I get to the Internet via CTSNet, a San Diego-based Internet access provider, and I use Mosaic, for a WEB browser. Say I want to ask BOL a question.
ACTION: ELAPSED TIME:
I click the icon on my Program Manager, computer dials and connects with CTSNet.......0:30
I click the World Wide Web icon, CTSNet’s home page appears....................1:56
I ask to go to http:/www.eb.com, BOL’s home page appears.......................2:50
I request and get the BOL search box...........3:50
It has taken almost 4 minutes just to get to the point where I can ask my question. I recently repeated this sequence ten times and found it took from 2:40 to almost 4:30 to get to the search box; the average was 3:25. And these elapsed times don’t reflect the fact that on three of my ten trials, I was cut off before I ever reached BOL and had to start over again.
But different gear may give different results. My neighbor, a computer engineer with a hot new setup (a Pentium 90 with 16 MB of RAM and a 14.4 modem, using Netscape), was able to get from his program manager to the BOL search box in about 1:15. It’s a lot less irritating to wait 1:15 to ask your electronic encyclopedia a question than it is to wait 3:25.
Back at my home computer, I asked BOL to find “Chicago,” then I clicked on the first article listed, scrolled to the bottom of it, and then went to the City of Chicago’s World Wide Web home page. The whole process took about two minutes. But once again, when I tried this on one of the Power Macintoshes at UCSD, it took only about 30 seconds.
Where’s the bottleneck on my home system? A number of Internet cognoscenti have told me it’s difficult to diagnose, at least for the moment. I assume that when I get a faster, more powerful computer and/or a better Internet provider and/or a better Web browser, I may well find BOL to be a joy to use. But I can’t imagine paying money for BOL until something has better lubricated my links to the Internet.