Sphinn - New Google Process for Detecting Near Duplicate Content

58
Sphinn It!

New Google Process for Detecting Near Duplicate Content

Went Hot: February 25, 2008 - 6:16 am
Posted By: SpostareDuro 497 days ago
Topic Type: News Story (Jump to http://www.seobythesea.com)
Category: SEO

A new patent application on near duplicate content from Google explores using a combination of document similarity techniques to keep searchers from finding redundant content in search results.

3 Comments

Who Desphunn This (0)

Comments

from billslawski 497 days ago #

Votes: 2 | Vote:

+ -

Thank you for the sphinn, Kimmie. I liked this patent application from Google a lot because it provided so many citations to other documents about duplicate and near duplicate content, and provided some good research about those other approaches.

from iamlost 497 days ago #

Votes: 0 | Vote:

+ -

This issue - and the SEs' solutions - will become increasingly critical. As duplicate filtering succeeds, and only one instance of that content is ever returned, determining/crediting the original takes on greater value importance. Unfortunately, the SEs do not much weight actual authorship unless hit over the head with a DMCA cudgel.

My auto-response 'Thank You' to Bill for continuing to bring us the future of search as described by patents - in plain language. As always (a couple of years == always in web time, right?) seobythesea sits atop my reading list.

from billslawski 497 days ago #

Votes: 0 | Vote:

+ -

You're welcome, iamlost.

Like many of the papers/patents about duplicates and near duplicates, this patent application doesn't talk about how a search engine might distinquish between an original author and someone copying content.

The only patent that I can recall from Google that did address that area was their patent on Agent Rank, which relied upon a digital signature like that used in Open ID to sign content, blog posts, blog comments, and other content.