Comparison of a Visual and a Textual Notation to Express Data Constraints in Aspect-Oriented Join Point Selections: A Controlled Experiment