IF the programmers had some way of creating a Google-internal-only search engine, that is, it works exactly like Google and I mean exactly, and yet can only be accessed from inside the Wikipedia frame, that could possibly work.
Is there some reason why the Foundation can't/won't buy a Google Search Appliance? (non-free content, etc?)