Globalization–ethics and plagiarism

This issue presents two exciting papers on different topics in software testing. The first, Optimizing Compilation with Preservation of Structural Code Coverage Metrics to Support Software Testing, by Kirner and Hass, presents results on mapping coverage computed at the source level to coverage at the executable level. Their suggestion is to modify compilers to add additional information to the executable version of the software to make it possible to back-calculate coverage measured on the executable source. (Recommended by Jose Maldonado.) The second, A Hitchhiker’s Guide to Statistical Tests for Assessing Randomized Algorithms in Software Engineering, by Arcuri and Briand, presents guidelines on using statistical tests in experiments involving randomized algorithms. The authors make the point that randomized algorithms have different characteristics than other kinds of experimental research, which means different statistical analysis techniques are needed. The paper analyzes several recent publications and recommends how their statistical tests could have been improved. (Recommended by Alexander Pretschner.) I wrote about The Globalization of Software Engineering in a previous editorial [1], followed up with a discussion of language skills to support globalization [2], uses of references and citations [3], and standards for research quality [4]. Another difficult difference that is affected by globalization is plagiarism. All journal editors that I ask agree that we detect more plagiarism than in the past. Part of the increase is simply that we now use technology to increase observability. STVR uses an automatic plagiarism tool that searches thousands (millions?) of documents looking for overlapping text. It reports the percentage of the text in the submitted paper that is identical to text in previously published papers. We can also view the papers that have the most overlap in a tool that highlights the overlapping text. As a result, we detect more plagiarized papers, and detect them sooner. STVR rejects 10 to 20 papers annually for plagiarism. Another reason for the increase is that many authors do not understand plagiarism. As I pointed out last month [4], scientists learn many things from their advisors, including what constitutes plagiarism. This, in turn, is certainly affected by culture. In countries with governments that are inherently corrupt, it is natural for plagiarism to be more common and accepted. In countries without a long tradition of individual ownership of ideas, plagiarism is not a natural concept. And in countries that are extremely competitive, the ‘anything to get ahead’ attitude can sometimes turn to plagiarism. Thus, the increasing globalization of software engineering certainly contributes to an increase in plagiarism. A prevailing problem seems to be ‘what is plagiarism?’ My Merriam-Webster dictionary defines plagiarize as ‘to use the words or ideas of another person as if they were your own words or ideas’ [5]. This seems simple and straightforward, but, as the saying goes, ‘the devil is in the details.’ In the succeeding text, I present several (anonymized) examples. Complete copying: Perhaps the most obvious and egregious case is when an entire paper is copied with only slight changes. I have seen this twice in my tenure at STVR. Most recently, the original paper was published in a fairly obscure outlet in 1978. By happenstance, the reviewing editor had read the original paper and found it, scanned it, and sent it to me. The new paper contained a few additional paragraphs and a few new references, but all results, figures, formulas, and text were copied verbatim. This was an easy case—the authors knew they were plagiarizing, did it intentionally, and added no value to the original paper. Our response was to inform their department chairs and Dean, and put them on a ‘no submission ever’ list for STVR. If they submit other papers to STVR, they will be desk-rejected without consideration.