Hypothesis: In the early days of the internet, people used the search engines their geek friends recommended, which meant that the best search engine (Google) became the most popular. Now computers are better established, and people use the search engines that come bundled with their computer/browser or the ones they're used to using. It'll only be smart geeks (who don't click on ads) that try out new search engines (just like they're the only ones that try out new browsers, like Chrome).
What's going to defeat Google at Search is when generic search engines become "good enough" and every website has an excellent integrated search engine. The competition is catching up faster than most people think. In the future nobody is going to ever type http://www.google.com into their web browser and Google knows this. That's why they're pushing into Mobile and Facebook so hard -- they're screwed if they don't.
This is a niche underneath two giants, but it's a sheltered niche. Google and Bing can't copy them without producing strange side effects for their normal users who make typos and copy/paste material into the search box. A slight degradation of service for normal users (who don't expect /map or /time to limit them to a single web site) is much worse than allowing a relatively tiny competitor to have an edge in features.
Google does this with name:value pairs, like site:yahoo.com or format:PDF. You are concentrating on syntax here rather than semantics: the same thing could be accomplished with the (poorly named) "slash" tag, letting users do slash:map, slash:time, etc.; obviously, you would then want to restructure the space of tags to make more sense.
I think people understand that text affects their search and punctuation (usually) doesn't. The "site" and "format" keys are at least self-describing in a way that "slash" isn't.
Back when I was a kid in High School, I was proud of the fact that I could enter enough words into a search string to find what I was looking for on AltaVista. I was guessing what words might appear in the article / paper / search result I intended to find usually, but it worked. When a professor at college first recommended Google I resisted changing because I felt like a pro at AltaVista, but in the end Google searches were just too easy to find what you were looking for. Long story long, the only way you're going to beat Google is making search more efficient than Google.
