IBM's Social Beehive and Discovery Search

By Clint Boulton  |  Posted 2008-01-21

IBM's Social Beehive and Discovery Search

Orlando, FLA. -- IBM served up a lot of technology for its 15th annual Lotusphere show here, but the uber vendor still has plenty of dishes cooking in its research lab oven.

Software engineers from around the world showed off, for the first time, several wares in demonstrations for the Armonk, N.Y., company's Innovation Lab Jan. 21.

One of the more interesting concepts is Beehive, an opt-in internal social network to help the thousands of IBM employees get to know each other better. As IBM's Facebook, Beehive has drawn more than 6,500 employees to the fold since hitting beta last October, allowing employees to share personal and professional information with colleagues.

Beehive has many of the characteristics of your typical social network -- profiles, lists, events, photos -- and users can post photos and videos like any other media-oriented site. The software also allows users to drag and drop information throughout their profiles.

Self-branding is the key with Beehive; users needn't provide required information the way other social sites do, Beth Brownholtz, software engineer of collaborative user experience for IBM Research, told eWEEK. Users can portray themselves the way they want, choosing what they want to share and what they want to keep private. Users can also leave comments.

Beehive currently offers search by people and by tags. One thing it doesn't do is rank the strength of connections, for example, to assemble contacts by frequency of contacts with other colleagues. This is considered an important tool in fine-tuning the efficiency of social networks in the enterprise.

IBM is using Beehive profiles to populate another important internal work-in-progress: Project Bluegrass. Bluegrass is a virtual reality application that lets software developers work on projects together, showing visual representations of ideas and Web data. The program attaches Beehive profiles to avatars and lets users hold virtual meetings, a useful proposition for a company whose employees are spread far and wide.

Social Discovery: Algorithmic Search with a Human Touch

Another application, Social Discovery, is IBM's internal iteration of social search, an increasingly popular method of giving algorithmic search a human touch.

Social Discovery is a search application that returns not only documents, but people related to a topic from the company's internal Blog Central application, any IBM Dogear tags associated with the theme, as well as user recommendations. Results gauge users' interests, user feedback about documents and the popularity of documents.

"First we add all the tags and bookmarked content to the document so that you will be able to find something by the words the user used, and we also take into consideration all of the activity around these documents," said Shila Ofek-Koifman, a manager in the information retrieval group for IBM's Research Lab in Haifa, Israel. Ofek-Koifman said Beehive is not currently a source Social Discovery can tap, but is a possibility for IBM in the future.

Also on the agenda for IBM:

- Chat Search, the ability to search instant messaging records. The technology scours Lotus SameTime messages from keywords, and will eventually let users find instant message records down to the minute.

- Project Jumbo aims to assist people with disabilities trying to communicate with colleagues.

Ira Forman, senior software engineer for IBM Research, has created Project Jumbo, an automated speech recognition software plug-in to help deaf people use Lotus Sametime by speaking into a transmitter. In a demo, Forman asked a computer program questions by speaking into a headset. Within three to five seconds, the program returned typed answers.

- Real-time Translation, created by David Brigida, executive project manager for IBM, translates text or speech in Lotus Sametime into nine other languages, including Spanish, Japanese, Korean and Mandarin.

The projects IBM researchers introduced do not currently have timelines to become products. Most of them are running on different schedules, and engineers could not comment on timelines for making the research a reality.

Rocket Fuel