Mostly the issues come down to the differences between searching and filtering and I put together the little example below which helped explain things;
Bob lives in a house. He's got lot of books one of which is the Great Gatsby. His bookshelf is in his lounge and everything in his life is nicely indexed and searchable and all result sets include the location.
Bob wants to find his book.
He does a search "book=great gatsby AND location=lounge". He gets the result that the book he's looking for is in the bedroom (where he left it).
Now if he'd used the same text as a filter then he'd have had no results as the book wasn't in the lounge and filters don't have the flexibility to, if no result matches exactly, to widen out to include "close enough"-matches that, in this case, give him the result he's looking for.This worked quite well so I thought I'd share it.
As a side note this also demonstrates the effectiveness of a really good search engine. Take, for instance, the majority of users habit of "filing" their emails into folders. If everyone in an enterprise was reliably able to search for things just think of all the time that could be saved by everyone just archiving emails out of their inbox rather than filing them away.
