1999 - 2001
– hypertext
Bowerbird Hypertext Search Engine
Bowerbird was a hypermedia theory web robot that was used before the age of Google. The robot semiautomatically crawled pages on the World Wide Web, indexing them against a set of filter words I had defined (words such as hypertext, hypermedia, multimedia, cybertext, etc), and building a public database of content.
Prior to Google any search of the web for something like "hypertext" generated literally millions of returns, the majority of which were about HTML code or basic web technical issues. The advantage of such a robot was that it provided access to otherwise disparate information sources in a single location, with some degree of 'editorialising' so that a search did not produce thousands of irrelevant results. It was, in effect, a specialised version of any major web search engine from the time.
Bowerbird had crawled over 350 web sites, 220,000 pages, and indexed approximately 10% of these (on the basis of relevance). It attracted around 300 unique hits a week, which is quite low for the quality of the material available.
Bowerbird was run on commercially available software (Phantom from Maxus). It was a good example of a simple, relatively low cost application of information technology to aid research. It took approximately five hours a week of maintenance, research, and updating.