When I reflect the current state of requirements engineering, then I recognize the following: There are a number of good tools for documenting requirements There are good and established methods for creating natural language requirements There are still relatively few standardized methods to properly handle recurring categories of requirements