Open source search result clustering
Posted on January 9th, 2009
Remember Northern Light? That was the search engine that did an amazing job of dynamically clustering search results into folders based on their topic or area. That same type of functionality is now open sourced at Carrot2. Jon Lehto, of Monster.com, who pointed this out to me, notes that Carrot2 ” builds categories from search results (configurable number), and the user doesn’t need to ‘own’ the search data.”
BTW, Northern Light is still around, although not for public Web searching. Instead “Northern Light provides strategic research portals to global organizations.”
Categories: everythingIsMiscellaneous, metadata
Northern Light, besides being my Dad’s favorite Canadian whiskey, had problems with finding relevant links. Insufficient data, properly displayed, is still insufficient data.
I have been, in spite of my skepticism of monopoly, been unable to shake the dreaded “Google Habit”. As The Software Piper of Hamlin said, “Follow the algorithm that works.”