Good article.  Very informative.  On performance, if I use a GUID as a Primary Key with a clustered index, won't that require the database to constantly 'reorder' the data since a clustered index determines how the data is stored physically.  Perhaps a larger fill factor is needed to help alleviate these concerns?

Example:  If I use an integer for my PK in the orders table, new orders are always appended to the end of the table.  Order 1,2,3,4,5 are all added in sequetially.  With a GUID, I have no idea where an inserted row will end up physically in the table.

Thanks for the info Andy!

Scott