Annus Mirabilis: 1993. An unexpected stimulus to ATG's efforts occurred in January 1993, when the Editorial Department at Britannica was approached by representatives of the University of Chicago, who proposed to help develop a LAN-based version of Britannica to run on the campus network. The University already had amassed considerable experience in dealing with large textual databases, notably the ARTFL archive of documents in French literature. Discussions continued for a few months, and a bit of prototyping was done, but in the end the proposal came to naught. The useful results of the episode were three: a demonstration in which some Britannica text on a university server was accessed via dialup connection from the EB Library and displayed on a PC in glorious green ASCII; credibility for the proposition that there might be a market for a networked Britannica product among colleges, libraries, and other institutions; and, in one of the earliest of the meetings, a mention of the word "Internet," heard then and there for the first time by the Britannica side.

From the outset Chicago and La Jolla took quite different approaches to conceiving Network Britannica. In Chicago it seemed natural to assume that the product would employ wholly proprietary software. In La Jolla the ideas of easy exchange of information, open standards, and off-the-shelf applications held great appeal. John Dimm, who had first-hand experience in developing the various flavors of the Compton's CMME and was therefore intimately familiar with both the difficulty of the platform-specific approach and the brittleness of the results, was especially ardent in championing the adoption (and adaptation as needed) of available and proven solutions. It is also worth remembering that even given development focused on LANs, provision had to be made for dialup connection as well, which at the time meant, at best, 9600 baud transfer rates.

The first concrete step in the direction of general solutions was the decision to abandon SmarTrieve in favor of the WAIS search engine. WAIS (Wide-Area Information Server) had been introduced two years earlier as the first full-text search engine capable of running over multiple, distributed databases. Contact was made with the WAIS office in Mountain View, California, and in a short time a formal agreement was reached for code-sharing. A WAIS staff programmer (Harry Morris) who was already at work on version 2 of WAIS was the main liaison with ATG, working chiefly with Brian Bartell and Amy Steier. Steier's work involved developing software able to recognize phrases in English text and treat them as fixed units for indexing and retrieval. Bartell sought to improve the estimation of relevance by the retrieval engine by incorporating into its logic a number of "experts" sensitive to certain conditions, such as the occurrence of a search term in a document title, or multiple occurrences within a document, or occurrences early rather than late in the document.

(A minor byproduct of this business association was an invitation to WAIS's founder, Brewster Kahle, to speak at the Britannica Editorial Convocation in October 1993, where Vince Star and his colleagues memorably appeared sporting propeller beanies.)

On April 14, 1993, the "eb.com" domain was registered with InterNIC, largely at Clarke's instance. Chicago was far from accepting the Internet as the platform for Network Britannica, but ATG was forging ahead, protected mainly by the 1,723 air miles separating them from the home office.

Up to the summer of 1993, Dimm was proposing that Adobe Acrobat be adopted as the document viewer for Network Britannica. The choice of Acrobat was in line with his view that established and widely used software was preferable to a proprietary solution. Acrobat, which provided page views of documents, would solve the problem of displaying formulae and the like, though not that of capturing them in the first place. As it happened, Britannica was at that point considering Adobe, among others, as a supplier of a new system to replace IPS. In the end, neither came to pass. Dimm gave a presentation at a computer science class at UCSD and there saw a demonstration of the World Wide Web, using Tim Berners-Lee's original browser. Shortly thereafter, in June 1993, he saw a pre-alpha-release version of Mosaic at the American Booksellers' Convention in Miami. He immediately began plumping for the Web as the platform and Mosaic as the viewer for Network Britannica. By the time Mosaic was first formally released, in version 0.5a in September, the challenges of data conversion and making WAIS Web- and Mosaic-compliant were already in course of solution.

A large sample of EB data was parsed (from BSTIF) into HTML in November and was browsable by means of Mosaic. By early December the whole article database was up and running. Bartell and McInerney had written a cgi (common gateway interface; a standard piece of coding later but a bit of rocket science at the time) program to link Mosaic, WAIS, and the document collection. (Special characters were still a problem; one exchange of emails concerned the all-important (for Britannica) "æ" character.) McInerney had built the necessary servers and firewall and had created a user-authentication system based on users' IP addresses. At a time when such was still possible, McInerney communicated directly with Marc Andreessen at the NCSA center in Champaign, Illinois, where Mosaic was under development, on getting the WAIS-HTML gateway to function properly.

The result was an application that invited the user to type in one or more search terms, which were then processed by WAIS in order to produce a relevance-ranked list of document references. These references were then presented in a Web page, created on the fly, with links directly to the documents, which were Britannica articles or portions of articles. In subsequent development this "hit-list" page featured a brief extract of each document (to enable the user to make a more informed selection), while the returned article showed the search term(s) in boldface.

In the second half of December the firewall was cautiously opened to a few people outside the ATG office, who were invited to look in on

Rik Belew was able to browse Britannica articles from his campus office, as was a key contact at the UCSD library, and Vince Star logged in from Chicago. The first semiformal demonstrations took place in what we may as well call the