Typical web crawlers operate at the page level--they simply traverse links, scanning and indexing entire pages of content. Oftentimes, though, the most important nuggets comprise only a small portion of the page. screen-scraper™ allows for granular elements in the page to be identified and indexed for what they are--publication dates are understood as publication dates and expense amounts are understood as expense amounts. When data is indexed in such a granular fashion more of the power of Oracle SES can be leveraged. For example, this allows for search results to be filtered in an intelligent fashion via the clustering capabilities of Oracle SES.