I have been thinking a lot about search lately, not only due to my reliance on Google that we all now share, but also because I currently reading an excellent book,

The Search by John Batelle

, about all things search and Google.

Therefore I was interested when a friend recommended that I check out a new type of search engine, based on the idea of clustering the results into logical groups. 

This search engine, Clusty, caught my interest partially due to the fact that only a few months ago we had a very interesting presentation at the Scottish Usability Professionals’ Association about creating usable search and knowledge management, and one of the presentations featured presenting the results in groups of similar themes. 

Furthermore, search, both within sites and across sites, is one of the most common topics of interest with User Vision and our clients.

Created by Vivisimo, a search software company from Pittsburgh, Clusty is a metasearch engine, retrieving results from Ask, Open Directory, Gigablast and others, including Wikipedia. 

The main engines, Google and Yahoo, are noticeable in their absence. Clusty may not be as fast in delivering the results as Google, but it does add an interesting new dimension to interpreting the results.  

In addition to the usual search results ranked by relevance, the left side of the results page features a set of expandable clusters created by indexing the site. The clusters help users see search results by topic so you can hone in on your search query and discover relationships between items. The clusters let you drill down by topic and find results that may be missed or were buried deep in the ranked list.

It also lets you create a tag cloud (a ‘Clusty cloud’) which visualises the strength of relationship to other topics using the familiar tag cloud display as seen in Flickr. You can click on any of the tags to go to Clusty’s search results for that topic. 

I’ve installed the toolbar on my browser and am trying it out as an interesting alternative to Google (which is still pretty faultless in my opinion). 

Have you tried Clusty or have any other search engine recommendations?