Collaborative Classifier from Verity
- By Jack Vaughan
Many a corporate Web portal has been tripped up by a need to classify content. Over the years, this has stubbornly remained a place where humans outshine machinery. In a new product that takes a new approach to classifying, search specialist Verity Inc. looks to better empower those humans in the organization that are most capable of classifying the content with which they work. It would seem to meet the needs of knowledge workers who say "No taxonomy without representation."
Verity has plenty of classification experience. "This whole notion of taking search and breaking it into compartments or 'Yahoo-style' categories has been compelling," said Scott Whitney, director of product management at Verity.
But to get it to work, it has too often required corporations to go out and hire someone trained in what Whitney calls "controlled vocabularies." Typically, that means someone with a Ph.D. in library science. And the number of categories in some areas of endeavor limits what these librarians can accomplish without adding resources. "This can be expensive," notes Whitney.
All of this drove Verity to build the Verity Collaborative Classifier (VCC) 1.0, itself built as a collaborative effort of sorts with input from Dupont, Raytheon and other end users. Released just this week, the classifier seeks to "let the people who use the technology, organize it," said Whitney. The operative principle that distinguishes Verity's approach is the idea of "applied workflow" that enables more of the taxonomy and classification management to be distributed to assigned subject matter experts within the organization.
VCC 1.0 supports various roles for maintaining taxonomies. While subject-matter experts can take on more responsibility, a corporate librarian still has a role in the process -- which leads one to wonder if this is just for the big organization.
In fact, capabilities in VCC 1.0 are said to lower the bar for small- to medium-sized businesses interested in better taxonomies for corporate information. The software allows users to import templates from ISO-2788, which is a standard way of representing thesauri in XML form. As well, a portion of the package allows VCC to mimic an existing portal as a taxonomy. "We also offer packaged taxonomies, including open-source taxonomies [for some key content areas]," noted Whitney.
The point of all this is for a company to know what it knows. "Managing taxonomies improves competitiveness," said Whitney. "You can discover work that was done in the past, and improve time to market."
Jack Vaughan is former Editor-at-Large at Application Development Trends magazine.