Store any type of data - SQL 2008
As you may have or may not seen from the initial documentation of SQL Server 2008 (formerly Katmai) product overview that few changes with regard to the data transition for the relational and non-relations quite good changes are introduced such as:
FileStream data to store/allow large binary data with transactional consistency, as of now the best practice is to keep away from binary data for the sake of performance & management.
Full-text search integration to make the searching facility for text and metadata with better text based indexes or on large text columns (personally I make limited use of fulltext, hence nothing much I can say).
Sparse columns to allow the storage of object models on to relational data with limited space costs, this is useful for content management application to make use of better SQL Server database.
Lastly, Large User-Defined types and aggregates mainly until SQL 2005 there is a limit for both user-defined data types and aggregation to not to exceed more than 8KB data limit. With the latest release of Orcas working with large UDF or UDA is not an issue as the system assemblies for Orcas uses mscorlib to support such data types data for SQL 2005 & 2008 version too.
So the upcoming CTP for SQL 2008 will bring more surprises for DBAs & Developers to make better usage of integrated products.
Also you can vouch or propose such feedback on Connect site.
Large user-defined types
Large user-defined types allows users to expand the size of defined data types by eliminating the 8‑KB limit.