Prototype
Page 2 of 8
Info-Strainer
Searching a conventional database is like using chopsticks to hunt for gold nuggets in a pile of ore: every pebble must be picked up and examined individually. Engineers at StreamLogic in Los Altos Hills, CA, have invented sifting software that lets informational “ore” pass but captures the gold-alerting users to interesting new content, such as news stories on the Web.
StreamLogic’s program monitors constantly changing content sources such as discussion groups, newswires and stock quotes and categorizes their information by topic, according to the frequency of certain words or word pairs. It then strains this categorized content through a mathematical filter; when content matching a preset pattern emerges, the system issues an alert or extracts the data in real time. A Web site on Middle Eastern politics, for example, could watch news feeds for stories containing the words “Arafat,” “Sharon” and “intifada”-and then present only those items to the site’s visitors.
0 comments. Share your thoughts » 0 comments about this story. Start the discussion »