Monday, February 16, 2004
If you do a search for "Eurekster suck" on Eurekster, it gets censored. It is most likely because of the pornography filter and not some semantic PR filtering tool because "Eurekster Blows" and "stuck without the t is censored on Eureskter" are both searches that are propogated to my friends.
Disclaimer: I do not think that Eurester blows or sucks. Eurekster's related searches, however, are probably accurate in saying that Rebecca Lord Sucks and ex-wives suck.
more on Eurekster's porn filter: last post on this page, first post on this page.
Overture finally updated their Search Term Suggestion Tool to have January's results. 405 people searched for Eurekster in January. 36 searches were performed for "eurekster suck". Interesting. There were no searches for "Eurekster Great". Very interesting how people respond to innovation- Change is Bad. As I posted previously, 32 searches were done in December for Eurekster.
It'll be interesting to see what February's searches will look like for Eurekster. I wonder if it'll have the same drop off that waypath shows...
Looks like Ebay is testing Vivisimo. Very cool. Now, I don't have to guess which camera brand is most popular, Vivisimo automatically clusters by brand and other criteria.
Update: Found out Vivisimo is using Ebay's API to deliver the results. Battelle ponders whether ebay will respond. This demonstrates the power of an API. Ebay gets free R&D, free clicks, etc.
In Wade Roush's article in MIT's Technology Review, Search Beyond Google, he doesn't mention Eurekster. However, he does touch on both dipsie's, mooter's and microsoft's search efforts.
As is described by Roush, Mooter uses user behavior similar to Eurekster to improve relevancy of search results. It also uses clustering technology to group results, similar to vivisimo.
Dipsie promises to index database driven sites more thoroughly and microsoft's development areas include natural language queries, indexing of pcs, and delivering search tools from other programs, both by users prompts and automatically during normal taks.
It'd be nice to know the differences between mooter's reranking algorithm and Eureskter's, which are both based on user behavior. I am looking forward to seeing dipsie, supposedly this summer.