Preliminary report on empirical study of repeated fragments in internal documentation

In this paper we present preliminary results of an empirical study, in which we used copy/paste detection (PMD CPD implementation) to search for repeating documentation fragments. The study was performed on 5 open source projects, including Java 8 SDK sources. The study shows that there are many occurrences of copy-pasting documentation fragments in the internal documentation, e.g., copy-pasted method parameter description. Besides these, many of the copy-pasted fragments express some domain or design concern, e.g., that the method is obsolete and deprecated. Therefore the study indicates that the cross-cutting concerns are present in the internal documentation in form of documentation phrases.