Sidebar: Had someone back about 1985 phrased his question, wittingly or not, “Has the Britannica been digitized yet?” the answer would have been “Yes, and in fact we’re already on our third computer system.” Publishing an annual revision of a 32,000-page encyclopedia is a major undertaking, and the development of computer-based systems for handling this huge amount of data promised similarly major economies. The text of the encyclopedia was first digitized – keyed into a digital database – about 1977; the data fed an RCA PhotoComp photocomposition device that produced the page film from which offset printing plates are created. In 1981 the data was ported to a more powerful Atex system, a system mainly used by newspapers because it offered a very robust set of editing tools. But when a full revision of the encyclopedia was undertaken – one in which every page in the set would have to be recomposed, rather than the 1,000 or 2,000 that had been typical for some years – Atex proved inadequate. It was replaced in 1984 by the Integrated Publishing System, or IPS, a purely software solution developed by the Watchtower publishing arm of the Jehovah’s Witnesses. Because Watchtower carried on a very heavy publishing program in 60-odd languages it had taken the initiative to develop a system with good editing tools and a WYSIWYG composition engine and one that natively handled a very large set of alphabetic and symbolic characters. Marketing, licensing, and support for IPS were handled for the Witnesses by IBM, on whose mainframe it was designed to run. Having installed IPS, Britannica’s Editorial Computer Services group undertook a program of customization, adding such features as hyphenation and justification, a spell checker, and various editorial management tools. The editing segment of IPS, called PSEdit, used a text markup system that, while proprietary, was sufficiently SGML-like that in time it proved readily convertible to that system or to HTML. (Less readily convertible was PSEdit’s use of color and reverse video as meaningful attributes.) A typical short Britannica article, displayed on an editing screen (an IBM 3079 terminal) looked like this:
One unique and essential bit of customization added by the ECS group was a method of inserting metatext tags, called hooks, into text articles and linking them in such fashion that they behaved analogously to what later came to be called hyperlinks. These tags were used in a variety of ways: to link illustration files and their related caption files to points in article text; to connect cross references with their referents; to connect groups of topically related articles with the outline of topics in Britannica called the Propaedia; and, most numerously, to connect terms and their citations in the Index with the actual points in text from which they had originated and to which they pointed back. In the graphic above, the pink strings are the hooks. The Index hooks and Propaedia hooks would later be exploited in Britannica Online to produce an elaborate navigation system over the whole body of 65,000 articles.