From NMI Wiki

MuseumsAndTheWeb: Auto Aggregating

The Walker is currently prototyping an "auto aggregator", basically a tool that would search various sites (google news, technorati, flickr, etc) for references to the Walker -- or whatever topic we choose.

The future of the web is mashups and open AP Is? and information sharing. We're trying to grab info and articles from as many places as we can and then run them through a republishing "pipeline":

1. Toss the spam
   1. Akismet has an open API to help catch spam blogs, etc -- and it learns over time
2. Try to categorize it
   1. Using a Bayesian filter -- similar to Akismet and other spam filters -- we will train the software to sort content into groups we define
3. Queue for moderation
   1. We will approve content by hand and correct (train) the grouping filter

Result?

We hope to have an at-a-glance look at every mention to the Walker on the web: via text, incoming link, or obtuse reference.

This set of content would be republished as an RSS feed and allow comments.