Digg Could Revolutionize Search Relevance

Users of Digg.com know how effective the process for promoting stories by votes or “diggs” has become; imagine if Digg was the way to get into a new search index with a Google-like algorithm to deliver results from it.

“What were those things?” Chia asked him.
“What things?”
“Like a business card, Crawled under the door. Then another thing, like a gray cut-out crab, that ate it.”
“An advertisement,” he decided, “and a sub-program that offered criticism.”
“It didn’t offer criticism; it ate it.”
“Perhaps the person who wrote the sub-program dislikes advertising. Many do. Or dislikes the advertiser. Political, aesthetic, personal reasons, all are possible.”

– Chia sees how Walled City deals with spam in Gibson’s Idoru.

Every time I use Digg, I’m able to forgive the relentless drumbeat of Web 2.0.

First a quick overview of Digg. The tech-oriented site takes user-submitted stories. Other registered users “digg” stories they like. Once a story exceeds a certain number of diggs, it gets promoted to the front page of Digg.com.

Users can comment on submitted stories as well, and they can report duplicate entries, spam, bad links, or simply deem it “old news” or “lame.” Enough negative reports will toss a story out of Digg.

Digg has its own search capabilities, and it works well enough at matching patterns. It’s not Google, or even MSN Search.

Back to the main idea now. Imagine if Digg could be the front end to an index. Every bit of content being fed to the index would have to be submitted by a registered user. Then vetted by at least a couple dozen different registered users.

From the search side, a well-developed algorithm would mbt men’s shoes handle queries, and be programmed to deliver as much relevance as possible. Every query would be hitting an index free of spam blogs and useless keyword-stuffed web sites. There would be people who would try to beat the system; there always are, really.

But instead of having to wait for the site’s administrators to update the index, users tabbing the content as spam would be doing so as they encounter bad results.

Rigorous enforcement of the terms of service for submitting to the index MBT Habari would be needed, with spammers being tossed as they reveal mbt habarithemselves. It might be the best combination of human and artificial intelligence possible for search, at least today.

Could it work? Maybe a better question would be “Why not try it and see?”