I tried to leave the following comment at The Digital Librarian, but there seems to be an error in their comment code - several tries and I couldn't get it to go. I also couldn't find any contact information, so I'm hoping posting here will eventually get back to the author. (c'mon, not even an "about me" link?)
Here's a fly in the ointment, paraphrased from a much more deep-thinking colleague.
The Schoogle federated search model works well (or will work well eventually) with providers of full text content because it's in their best interest to get their content before the eyes of the general public so the general public will buy their articles.
But what incentive is there for the producer of a strictly A&I database to open its content to Schoogle? Right now these databases, say Biological Abstracts for instance, make their money by selling their content to libraries. What income would they derive if they gave it all away to Schoogle?
You ask, "If Google becomes a better provider of scholarly articles and information than a typical university library...", but Google may not ever become a more comprehensive provider. Right now they sure lead the way for convenience, which goes a long way for the typical researcher, but I'm not sure they'll ever be the most comprehensive. Could be wrong, that's just my opinion.
Having said that, I have to agree that I have yet to see a federated search tool that a typical student would choose to use before Schoogle!