Kewl !! “Wiki-Based Search Engine Claims To Be Web’s Biggest” ( via URLwire ) . The real cool thing in it is that it enables user to edit search results. Wow. This is significantly differs from google [which uses pagerank( hyperlinks )] , it is a built a huge whois kind a database of all websites and creates index of summary of the websites thus having a greater coverage of the web.
I think it is still a very monolithic design, what would be a still cooler is to create a crawler for each user doing search with each crawler having a kind of balance between the index dataset duplication vs not requiring to ping a lot other nodes before finding the required index for search query. ( Basically a P2P kind of model , with each node as a crawler and each crawler having not much of duplication but requiring few pings to find its results ).
This way each crawler can index more elaborate summaries and solve the depth problem that it is faces. Also here the user only edits his own search results. What one person is looking for as results in a search term is not same as another person’s expectation of the same search term.
mtrajan.com 