A "White Paper" is a formal recommendation for a new product, or substantial changes to existing products. They are written over a period of years and formally presented at SHARE meetings. This particular paper was sponsored by the SHARE database project (although in fact it was the work of a single individual - as such I expect IBM will take it less seriously).

The White Paper pointed out a number of deficiencies in the current IBM DBMS product line, and suggested gross improvements. Among the observations:

• Current products are not reliable enough. Software failures are not acceptable. In the event of a failure, the future DBMS should automatically recover without human intervention.
• Continuous operation is critical. In support of continuous operation, the future DBMS must support online utilities, and allow an installation to update resource tables without interruption of service.
• Unattended operation is important. Current DBMS products require a master console operator.
• Distributed data requirements: data must be available from any system transparently, with no external evidence of network navigation. The network of DBMSs should put forward a "single system" image. The future DBMS should provide the ability to replicate data at two sites for purposes of redundancy and improved performance, automatically synchronizing them in real-time.