Interactive search refinement

Today Altavista rolled out a shiny new query refinement system, codenamed ‘Prisma.’ By giving users an assortment of possible query expansions, they’re hoping to close the search loop which usually ends either in success or complete failure:

AltaVista Prisma offers users a wise array of closely related words, phrases, concepts and/or names, to help users rapidly locate needed information and direct their search experience. While some tools focus on matching searches with general, high level directory categories, and others cluster results to aid in query refinement, AltaVista Prisma allows users to do both. Additionally, it provides parallel terms to help point users toward associated topics of interest.

The best part of the system is a simple interface widget: to the right of every additional query term is an extra button (>>) which replaces the current query with the given word. Most query expansion systems I have used in the past have suffered from the common malady of pigeonholing: as soon as supplementary terms are added, the query quickly becomes too specific to find the desired page(s). From a bit of brief interaction, the query replacement button seems to aleviate this problem (when coupled with the right terms)

A search for blogdex returns a host of possible expansions, including bloggers, MIT Media Lab, Project Info, and Weblog. Clicking on MIT Media Lab reduces the number of results from 8,075 to 49, but retains a list of parallel and higher order extension terms.

Anyone who was around the web circa 1998 remembers Altavista. In a hot market for search users, “the search company” was constantly innovating: first with image search, first with translation, first with filters, etc. But they had no conception that such a simple idea as PageRank would upet the market as much it did. Well I’m ready for some more competition.. if Altavista would just get rid of those damned pop-under ads, I’d use them more frequently.

One thought on “Interactive search refinement

  1. Hmmm, I remember Metacrawler having a similar “search subset” interface, though that was based more on meta tags, and we know how reliable those can be. 😉

Leave a Reply to Paulo Cancel reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s