What is the history of the Knowledge Base?

Indiana University's UITS Support Center has been answering computing questions via email, telephone, and walk-in since the early 1980s. Even in those early days, constant change and repetitive answers were the modus operandi. These two conditions encouraged staff to maintain an official, correct source for frequently changing and frequently requested information. Consultants also needed a repository for results of difficult or lengthy research they had undertaken to resolve some problems.

The first online "Knowledge Base", established in about 1988, was a directory of files on a shared VMScluster account, known as the "M" directory (named for Marie Meyer, one of the originators). Consultants answering questions via email found it convenient to base their responses on these files rather than repeatedly writing the answer from scratch. Upon sending information to answer an original question, they also created a more general version of the answer to include in the M directory for reuse.

This directory quickly grew into over a hundred files, often with multiple versions. There were no standards for the format of the files, and answers were written, revised, and maintained ad hoc. In about 1989, Support Center staff attempted a more logical organization of these files in a VAX Notes conference, called the "Pat Answer Library", but because of the unwieldy character of VAX Notes, this never really caught on.

By 1990, the Support Center was determined to find software appropriate for storing, maintaining, and retrieving this Knowledge Base, and examined several text-oriented, LAN-based databases (LAN-based so that the files could be shared, yet independent of any central operating system). The staff settled on "IZE," a commercial product that ran only on DOS, a major limitation. Furthermore, budget constraints allowed only a site license for five concurrent users. Nonetheless, IZE served well for almost two years. It offered content-based outlining and keyword searches that made maintenance easy. The Knowledge Base grew to several hundred files written by Support Center consultants or adapted from FAQs found elsewhere on the Internet.

Then came Gopher. In the early 1990s, a clever, ambitious young staffer, Scott Hutton, created a full-text searchable, world-accessible Gopher server that revolutionized the Knowledge Base. Support Center staff stopped thinking of the Knowledge Base as a tool for their own private use, and started orienting information for public use. Gopher allowed around-the-clock public access to the Support Center's FAQ repository.

The staff had been enchanted by Gopher for about six months when they first heard about the World Wide Web. Scott had the Knowledge Base on the web within a few months. Almost immediately, the Gopher version became an albatross. Scott wrote many hacks to make the Gopher version simulate the web (e.g., linking texts), but the team was, in effect, maintaining two copies of the Knowledge Base. The simplicity of the Gopher version handicapped the web version. Staff could not write texts embedding hotlinks or internal references. In March 1995, the team disabled the Gopher version of the Knowledge Base, freeing them to exploit HTML features to their fullest.

During the winter of 1995-96, the staff rewrote the Knowledge Base software to provide easier maintenance, faster searching, more search options, and other features not possible when the original web version was written. KB2, written primarily by Matt Liggett, with additional modules written by Alan Meiss and John Nienart, replaced the original Knowledge Base on March 6, 1996.

After over three years of work by Mike MacKenzie, Ed Dragomer, David Jantzen, Mark Meiss, Mike Fragassi, Josh Reedy, Andrea Donderi, Jonathan Phillips, and others, a new iteration of the Knowledge Base code, KB3, went into production in July 2002.

The Knowledge Base received more than 15 million hits in 2008, and currently contains more than 14,500 documents (slightly over 6,000 are active). Support Center consultants still use the Knowledge Base to capture solutions to problems users experience. Knowledge Management staff work with approximately 80 service providers and subject experts from departments and units across all eight campuses to identify, collect, deliver, and maintain both internal and external content in the Knowledge Base repository. This content functions as source material reused in other online information services and help systems, including most of the content in the UITS web space (http://uits.iu.edu), the TeraGrid Knowledge Base, and the Sakai Knowledge Base.

Comments/Questions/Corrections