Some people used to believe that data warehouses would quickly replace databases for both online transaction processing (OLTP) and online analytical processing (OLAP).
Of course, they were wrong. Why can data warehouses not replace databases and become "operational data warehouses"?
How radically would data warehouses (and their data-mining tools) have to change to become a viable replacement for databases?
Would they then essentially become databases that simply supported OLAP? Why or why not?