Recently had a great experience using {rlang} in #RStats. I had over 50 equations for scoring cognitive tests that needed checking across hundreds of participants. Instead of writing checks for each one, parse_expr + eval_tidy let me map through equations and flag discrepancies. A simple example:
