"The key value of this integration is that you don't have to go out to another system to do your search," said Aaron Brown, IBM's program director for content discovery and search, in an interview. "As there's more and more collaborative information being created, people need to be able to get to it more effectively."
Lotus collaboration products -- Notes included -- and WebSphere each have built-in search functionality, but their respective search engines index only data found within the products themselves. OmniFind Enterprise Edition indexes more than 30 types of enterprise content and 200 file types, including Web sites, file systems, enterprise content management repositories, databases, and SharePoint.
The new version of OmniFind Enterprise Edition also includes some limited semantic search capabilities, aiming to pull meaning from search queries and results. It used to be that when someone searched for "Joe's phone number," only results containing those exact words would be returned. Now, however, OmniFind Enterprise Edition understands that search is actually seeking an extension or 10-digit phone number of someone named Joe, and searches the company directory and documents for any such information.
That capability is part of Unstructured Information Management Architecture, or UIMA, an open framework for embedding text processing and text mining capabilities into products like OmniFind. Though OmniFind Enterprise Edition, as sold, stops at abilities like finding names and phone numbers, UIMA can be employed for such complex tasks as analyzing whether a document expresses a positive or negative sentiment. The new version of OmniFind Enterprise Edition makes it easier for businesses to access and integrate this framework into the product.
IBM promises to integrate more and more capability into search. For example, the new Lotus Connections social networking product includes a tagging and social bookmarking feature called Dogear, which IBM has already integrated into its own internal search engine. IBM's Brown also points to visualization technologies under development at IBM Research that could return a search query with a visual data set instead of only a list of results.