While we used Scraper (prior to the first public release), sometimes we had some problems determining the structure of the template using only tags. So, we started to take into account the content of tags on the template.
And will capture "something"
But won't match:
This is quite useful on tableful pages that it's hard to determine a unique structure.