Tim Almond

I’ve been thinking some time about whether there are opportunities for having a tool which does "Bayesian Searching", in a similar way to spam mail filtering in Paul Graham’s A Plan For Spam.

A browser toolbar could allow a user to mark pages that were not good matches, particularly pages which try and defeat Google’s indexing (SEO Spamming). It could then count good/bad words and as searches are down, exclude pages from the search results which have a poor good/bad ratio.

Alternatively, a little manual inspection of the words might give options for search terms to exclude from searches.

Implementation-wise, I think that it would be best done as Mozilla XUL with communication with the Google API and locally, a web service that would communicate with a lightweight database to store the results.

—–

Tags: Ideas //