my first theory was andrew was in the history of the document and google had an old version, but no.

my second theory, as there is a french version of the andy andy article, maybe french to english convertion of andy = andrew, but according to babelfish, it’s just andy. so no.

Third theory was maybe wiki is doing cloaking to the googlebot, but wget says no to that idea.

Why not link to wiki/Andrew_Snoid ? If it was wiki/Andy blah would it still get there ?

why does info james point to wiki/James_(Nip/Tuck) and not wiki/Jimmy_Jimmy ?

Weird behaviour, now i’m confused and wont sleep properly tonight Its definately a different smell to the normal google indexing, esp as site:en.wikipedia.org andy doesn’t list andy andy in the first page.

Even weirder, no other articles point to the Andy Andy, so googlebot would have to come in from the all pages index.

All the nationalities I tried (Australian, Swedish, Dutch) map to the wiki page about the relevant country. Some have additional titles that map to the country article (so /wiki/Australian goes straight to /wiki/Australia), but others, like Swedish, have a disambiguation page that, by my reckoning, is ignored. I suspect that the Spanish -> Spain style mapping could be due to simple word lookup. It’s not simply word stemming, since you’re never going to get from Dutch -> Netherlands by that route.

The Microsoft example is very puzzling to me. Searching wikipedia using google with the query ‘microsoft’ returns the obvious /wiki/Microsoft page. The Game Studios page doesn’t appear in the first 10 links. The /wiki/Microsoft page is both a better answer, and seemingly simpler to identify. I can’t think of a reason that the Microsoft_Game_Studios page would be returned instead. Any ideas?

Another weird example is searching for ‘Andrew info’. The OneBox result is wiki/Andy_Andy. That article isn’t linked anywhere I can find, and I can’t understand how it would be the chosen article. Again, there’s obviously an Andrew->Andy mapping, but I can’t see where this is being made using wikipedia data, so perhaps it’s being done separately using Google resources?

any thoughts about;
microsoft info goes to /wiki/Microsoft_Game_Studios even though /wiki/Microsoft is a thorough (and more applicable page)

spanish info goes to /wiki/spain

and ‘java bad performance info’ doesn’t return anything, is google’s indexing so good it doesn’t bother to waste bytes returning common knowledge ? thats clever