Great work. This is similar to the Piggy Bank/ Solvent project. However, it is more advanced and the browser automatically suggests appropriate mashup scrapers. The public mass can create wrappers for different types of web pages and make the browser intelligent collaboratively so that the browser will gradually understand the data in pages. This is aimed at slowly turning the current web into a Semantic Web using wisdom of the mass. But surely there are many difficulties due to the various unpredictable forms of structured and unstructured data in the current web.
Archive for mashup
Looks like a great place to find mashups and APIs.
This is about making mashups. It also uses Solvent for wrapper creation.