Thanks for the comment Alexander. Obviously, the selected compute engines should support the underlying open table format, so cross compute engine interoperability can work. Same for metadata, governance and security.