The issue focused on language features and testing without proof of correctness and conceptually sounds pretty solid: http://www.cs.syr.edu/~meijer/introduction/language_features.html <http://www.cs.syr.edu/~meijer/introduction/language_features.html> Assessing cross-language interaction can be challenging and error prone. What about allowing users to construct language features as needed? What about teaching users how to determine what features they