An empirical study of the textual similarity between source code and source code summaries