ZDNet writes about Swicki tapping communities for search and really describes a new way of making search more relevant. This is the question most search engines are currently trying to solve and the model of Swicki seems to be similar to Rollyo:

Rollyo offers the ability to search the content of a list of specified websites, allowing you to narrow down the results to pages from websites that you already know and trust.“

… but then again, not quite. Since they don’t only allow for predefined filters, but also measure user behaviour to identify which results will be relevant in the future. This also what might differentiate them from Google: it’s not just about what people „voted for“ by placing a link but also about what they actually visited.

Swickis combine Web crawling with filters defined by site owners and algorithms that analyze user behavior (keywords and pages accessed) anonymously and automatically, re-ranking results based on the community’s search actions.

(Just as I typed this I thought: who knows, if Google isn’t already measuring our behaviour anyway? – I mean, how would we be able to tell? In theory, they can measure our clicks on the pages of the search results – but in addition they can track us on any page that has Google Adsense Advertising – which would mean a lot of pages across the hit- and niche-websites of the web.)

Seems to be an interesting tool – if I have some time over the weekend, I might start my own Swicki search in this blog…