A mini-workshop with representatives from the data-driven science and database research communities was organized in response to suggestions at the first XLDB Workshop. The goal was to develop common requirements and primitives for a next-generation database management system that scientists would use, including those from high-energy physics, astronomy, biology, geoscience and fusion, in order to stimulate research and advance technology. These requirements were thought by the database researchers to be novel and unlikely to be fully met by current commercial vendors. The two groups accordingly decided to explore building a new open source DBMS. This paper is the final report of the discussions and activities at the workshop