Bruce Schneier, renowned cryptography and security expert, has had personal experience of plagiarism, by three Pakistani academics in, of all things, an ACM publication. The reaction of ACM – seemingly, they have no policy about plagiarism and therefore cannot ban the authors from submitting again – seems rather strange to me.
I have some experience with plagiarism, though mostly by students. When teaching at night school some years ago, I caught half a class in plagiarizing, with some students expelled as a result – and I wrote a letter to CACM about the experience. In those days I had to sift through Lexis/Nexis to find the copied text.
Times have changed. At the Norwegian School of Management, we have started to run student papers through a tool called SafeAssignment which seems to work well – it gives me as a teacher a score for plagiarism for each paper submitted (matched against anything on the Web + all papers previously submitted to SafeAssignment). The match is “fuzzy”, and catches even those students who thought they could get things in by changing a just a few words.
Most of the effect is probably preventive – students refrain from plagiarizing for fear of being caught. But I have caught students plagiarizing with this tool and have also used it for control of suspicious papers. It is much less work-intensive than the “pick a prominent sentence and Google it” technique, and seems to catch most copying. Of course, you have to manually check every suspicious paper, since the score can be driven up by quite legitimate quoting.
But why stop at students? I think (as an ACM member and author) that ACM should institute as a policy that they run all received submissions through a tool like Turnitin or SafeAssignment – this will immediately catch flagrant plagiarism such as the one reported by Bruce. In fact, it would be a fun excercise to run all ACMs published papers through a plagarism tool – or, why not, every academic paper ever published….
Google’s strategy is to “organize the world’s information”. Eliminating plagiarism would be an interesting advancement of that strategy – after all, plagiarized material does not really add information, except about the plagiarizer. In the long run, we should have everything back to first normal form, with no copies – just links…..
(And yes, parts of this text is plagiarized from my own comment in Bruce’s blog. At least I acknowledge it, and have extended it….)