Heard in the aisles
Attending our annual conference in San Jose is always refreshing—I get a chance to see professional friends and colleagues and discover what’s really on people’s minds. Often, the most illuminating times are those during the breaks between conference sessions, when I can surreptitiously listen into conversations.
I was more than a little surprised to discover that users are still incredibly frustrated with—and uninformed about—the state of enterprise search. We have reported on the technology extensively—and continue to do so. And, our parent company has spun off enterprisesearchcenter.com, as well as an annual publication (The Enterprise Search Sourcebook) and the Enterprise Search Summit, which is twice a year (in New York and San Jose).
Search is arguably the single most important component of knowledge management (an attitude, not an application). Regular readers of KMWorld learn about the rise (and fall) of search vendors, and we’ve reported upon search as a platform for developing precise applications
Most recently, we’ve witnessed the successful blend of search and business intelligence.
Users continue to bring the expectation of Web search into the enterprise, all the while realizing Web search is a different, simpler, animal. And why not? Plenty of interesting work continues to go on in the Web search world. Take OrcaTec, for example, whose Truevert "solution" is built on the concept of semantic search (another way of saying based on the meaning of the words), according to company founder Herbert Roitblat. The usual approach is to use human knowledge engineers to construct an ontology or related structure that tells the computer which words are related to which other words and how they are related. He claims the semantic Web is a good example of that approach, but the approach doesn’t serve the search world because it relies on words fitting into categories that are stable from time to time and from person to person. It requires someone to actively and accurately tag each document. With Truevert, Roitblat says, OrcaTec takes the approach that meaning derives from the pattern of word usage in the language. (You may want to read more about these ideas at truevert.blogspot.com.)
Many bright people are banking that their approach will solve the problems associated with enterprise search—and more are entering the fray all the time. Just this week, I got news of ZeBAze, which the developer calls a new generation of Windows search engines powered by artificial intelligence functions and designed to mimic the human capability to use flexibility during the search process. I discussed that approach with an analyst, who said there has been a lot of research into it—variations of which are already in practice. Only time and, of course, the market, will tell how well that company executes its vision.
It only makes sense that as users’ requirements grow, so do their expectations, and, I guess, their dissatisfaction. So, after all, what I kept overhearing just makes sense. The good thing is the very best companies continue to innovate, and intriguing new approaches emerge.