QIP = PSPACE. Either that's pretty cool, or incomprehensible math speak. :).

Edit: QIP and IP are related to theorem proving. Of course, the article uses "a problem instance is in the language", but if you consider theorems to be an exercise in string rewriting according to a grammar(which contains both the inference rules and axioms of the logic), that all makes sense. All "true" decidable statements are can be "parsed" using that grammar.

Now, I freely admit, I know less than some of this may imply, but hey, its a lot of fun.