In security too often services are understood not from first principles but via characteristic mechanisms used for their delivery. Access control had got tied up with DAC, MAC, RBAC and reference monitors. With developments in distributed systems security and with the opening of the Internet for commercial use new classes of access control mechanisms became relevant that did not fit into the established mold. Trust Management was coined as a term unifying the discussion of those mechanisms. We view trust as a placeholder that had its use in driving this research agenda, but argue that trust is so overloaded that it is now an impediment for further progress. Our petition asks for a return to access control and proposes a new framework for structuring investigations in this area.