Most of the times, behind a web search, there are three motivations*:
1. you want to find a specific site. For example, you want to find the official Google blog or the MIT site. If you knew the address of the site, you would type it in the address bar. So it's natural to type the name of the site in the address bar. Many browsers include Google's I'm Feeling Lucky (or Browse by Name, a variation that redirects to the first result only if it definitely answer your query).
This feature is already available in Firefox, Google Toolbar for Internet Explorer and could be added if you create a new search in Opera, using this address: http://www.google.com/search?btnI&q=%s .
2. you want to find information. If you want an overview of a subject, Wikipedia usually provides good content, so you might add wiki (or wikipedia) to your query. A good place for an information guide is Ask.com, a search engine that lets you expand or narrow your query. If you don't know the vocabulary of a domain, Ask.com gives you some tips.
Google offers some Onebox results for simple answers like "China population", and Ask.com expands the concept for general queries like China.
3. you want to make a transaction (e.g.: buy something). This is the commercial part of a search engine, the place for ads, Amazon, eBay, Froogle (and Froogle 2.0), Bizrate.com, Dealtime.com.
The problem with transactions is that often you want information about a product before buying it, you want unbiased reviews, comparisons. And many search engines don't understand that.
Yahoo Mindset doesn't try to guess your intention, but it lets you set the right balance between information and transactions.
This relation between address books, encyclopedias and shopping carts is the heart of a search engine and the way it handles this relation influences.
To sum up, the unclear intention behind a query causes mixed search results, so it's a good idea to state your intention first and use the right tools.
* Andrei Broder, A taxonomy of web search, 2002 [PDF]