I'd like to hear an explanation of what the problem is with a clustered index on a uniqueidentifier. Is it just the fact that as the table grows, the index will need to be reorganized often, or is there a more fundamental problem with width or with the datatype itself?

I have some tables with clustered indexes on uniqueidentifiers, but they are relatively small lookup tables that rarely change, so I haven't been concerned about them up to now; are you saying I should be?