This article was clear, concise, and well written, and it will be a big help when we do make the transition to SQL Server 2008.

I am curious about the performance testing you listed though. Have you tried any time tests on similar datasets to get actual time results? Could your choice of indexes have had a major impact on it?