I just wanted to throw in my two cents regarding the search engine.
Using Google might be a good temporary solution, but I think using it permanently, or even semi-permanently, would be a bad idea. From an editing standpoint, as Erik Moeller rightly pointed out, it can lead to duplicate articles that aren't found for weeks or even months, if ever.
From a user standpoint, it's even worse. An example is [[Mexican Cave Fish]]; created over a year ago, it doesn't show up in Google at all, either under that name or under its current title [[Astyanax mexicanus]]. (As an aside, that page probably needs to be moved by a sysop.) There is no way of easily telling how many pages are in a similar situation.
A search function is essential; this thing doesn't have a conventional index. I strongly feel that Google is not up to the task. I completely understand the problems with server load, and hope we can get that fixed in a more conventional manner soon (if another machine in the rack is what's needed, I'd certainly contribute toward that, as I'm sure would others). However, I'm wondering if a more efficient short-term remedy might be to find some other function, one less integral to the use of the Wikipedia, to disable instead of the search function.
Hephaestos