Tripp presents a fairly balanced reaction to the new technology, and stresses a message I've been trying to hit home recently: Given the massive number of changes coming in SQL Server 2005, DBAs must learn all aspects of SQL Server, at least enough to veto bad ideas from the development side. DBAs cannot rest on these changes. There is too much; and developers, unlike DBAs, get excited and want to implement now, not wait until a technology is proven. DBAs will need to know how to properly argue when the time comes.

Tripp also points out that set-based T-SQL is still the clear winner in most cases. T-SQL has been expanded with lots of new features that will help get rid of cursors and loops: Recursive CTEs, row numbering, and the APPLY operator are some of the more important ones on the list. Before you consider migration to SQL CLR, look at these new features. They may solve whatever problem you're having better and in a more performant fashion than CLR methods will.