Open Source Fuels IBM, Yahoo Enterprise Search Engine

Designed for use by businesses and other organizations, the no-charge application is installed on a server and provides access for a number of users within the enterprise or for those accessing the site.

The major changes came as a result of user feedback IBM received. About half of the 22 resolved bugs were reported by users via IBM OnmiFind Yahoo Edition online community.

"We built a Web community where users engage in dialogue with our developers and our team on how they use the product," Aaron Brown, program director for IBM search and content, told LinuxInsider. "So, from that we were able to understand the top gaps in the product, capabilities we didn't provide in version 1 that customers really wanted."

Original Edition

The original version of IBM OmniFind Yahoo Edition was launched in December 2006. The release was an aggressive move to change the landscape of the search market and offer an enterprise search product at no charge, Brown noted.

IBM OmniFind Yahoo Edition indexes up to 500,000 documents per server in more than 200 file types, including Adobe (Nasdaq: ADBE) PDFs and Microsoft (Nasdaq: MSFT) Office Word and Excel files, in more than 30 languages.

"Our goal at the beginning was to expand the market for search and make it much easier for every organization, be it large or small, to get started working with search and have a world-class enterprise search solution they could deploy without any upfront costs," Brown explained.

Competitor Google Mini, which supports 220 different file formats, comes in multiple versions. To search up to 50,000 documents, including all hardware, software and one-year of support, Google charges US$1,995. Up to 100,000 documents will run $2,995 and 300,000 documents will cost $8,995 for one year.

IBM has been in the enterprise search arena for 20 years, Brown noted. While the company was able to bring its experience and technology to bear in creating OmniFind, it is available at no charge in part because at the heart of the search engine is the Lucene open source indexing library.

"[It] has been pretty actively developed by the community, so we were able to leverage a lot of the work the community has done there, but we also added quite a bit of new innovation around that ourselves," he stated.

"We've been working in search and a lot of enterprise search technology. We took pieces of that technology that exists in all of our higher-end offerings and brought that into [OmniFind]. So, it's not purely and open source product. It has a lot of our own innovation there to," Brown added.

What's New

In the four months since the release of IBM OmniFind Yahoo, some 16,000 organizations have downloaded and are using the software, according to Brown. About 800 users have joined the online OmniFind community and have provided valuable feedback in terms of how they used the application and what needed improvement.

"The release of the product we announced today really is the culmination of all that work with our customers and with the online community that has built-up around this product since its launch in December," he said.

More than half of downloads for OmniFind have come from users outside the U.S., Brown stated. Although the product has always been available to search in more than 30 languages, all the interfaces and administrative tools and documentation were only available in English. With this release, IBM has translated those crucial elements into 14 new languages. Users and administrators can make the switch without pressing a button. Using the browser's language preference, the administration application and the search application automatically translate any text.

"We translated it into 14 of the worldwide languages where we saw the most interest for [OmniFind]," Brown said. "Now people in Europe and Asia and really all over the world can use OmniFind Yahoo Enterpise Edition in their native language without any additional work."

Some other gaps that the release addresses include the ability to conduct more sophisticated searches. IBM added advanced search capabilities that allow customers search and filter their results based on metadata. For example, users can now search Web pages authored by a specific person.

The company said it has also made it easier for organizations to embed OmniFind into other applications and solutions. Version 8.4 contained a powerful graphical tool for customizing the interface to match, for example, a corporate logo or corporate colors. Users could alter the look by simply pointing and clicking. Version 8.4.1 has extended the point-and-click customization to include every component on the screen.