In the Find Stories view, the user should be able to do a blanket raw-text search of stories having paragraphs matching given text. Nothing special here--no Lucene search indexes or anything like that, just old fashioned mySQL lookup as in [SQL pseudocode follows ..] "SELECT .. FROM .. WHERE paragraph LIKE '%" + replace(searchText, "'", "''") + "%'.
Yeah, this means finding any HTML that the editor might've injected into a paragraph.
I'm asking for this because we've already got an XML export that we load into Lucene that we query against but Lucene (our flavor being Lucene.NET) doesn't handle certain queries like "*@primedia.com" and pukes all over us. It's really rather gross. :P