A number of folks have asked about this in our forum so I thought I would share my response

The old Database Project that exists in Visual Studio 2005 are really an ad-hoc script storage project type so that users can quickly create ad-hoc scripts or queries to run against the live database and subsequently be able to check those scripts into a Source code control system. The old Database projects also allow for storage of object definition files but the emphasis are on single object support which are deployable via connecting to the target database and executing/running that script against that database.  These projects are more geared towards very quick ad hoc development to compliment the VB, C# data tier development in regards to datasets etc.

The Team System Data Database Project are alot alot more richer in support scenarios and features. Team System Data is all about creating a managed team environment so that enterprise database developers and DBA can managed the changes that they want to make to their different database environments. In brief TS Data Database Project provides the following rich feature set which does not exist in the old database projects

- A very rich database change management environment for creating offline, file based database schemas and objects which have an associated test environment<with test data, and actual test cases to verify any database structural changes>. These projects can be placed under version control.

- A complete offline database development experience where each object is understood in relation to all other objects in the project system from a semantic and syntax perspective. Thus allowing for detailed reporting on all objects effected if a single rename of an object occurred.

- The ability to compare all changes made to your local database project versus what’s is on your live product database prior to deployment and finally the ability to create a customizable deployable unit which can be run from directly within the project itself or via SQLCMD or from within SQL Server Management