"Reproducing results" in a scientific context doesn't mean taking the original author's data and going through the same process. It usually means taking different data and going through the same process. Having code on hand to put new data through the same process makes that a lot easier.