The point of this article is that scholars often make research contributions that do not fit neatly into the category of peer-reviewed journal papers, and that impacts the advance of knowledge negatively. First, the products of research simply do not always fit into the mold of peer-reviewed papers. Second, articles are often protected by a paywall, which restricts access in a number of ways, for example, one must be a member or subscriber. One way to make papers and other products more accessible is to follow an open-source model, which solves the accessibility problem but makes it difficult for scholars to get credit for their products even if they have been widely adopted. To address these problems, the authors propose the use of reuse graphs that provide a visual representation of artifact reuse along with reuse metrics.
While there is merit in the authors’ argument in favor of reuse graphs, the problems lie in the vagueness of the article. Ostensibly the article is about the reuse of research artifacts, but neither is defined.
If a software engineer writes a Python function that speeds up machine learning, and thousands of artificial intelligence (AI) researchers begin using it, that is a contribution. But if somebody reuses paragraphs from a research paper, that is plagiarism.
Furthermore, there is no definition for artifact. Most people would agree that papers and software could count as research artifacts, but what about a PowerPoint presentation or a template for a website? This is unclear.
This is the beginning of a conversation, albeit from an odd entry point, on important topics much bigger than the issues addressed in the article. What counts as research? How is it measured? If wide adoption is the key, then do memes count as research? Are research output and impact the most important measures of scholarly success? And so on. This article begins a discussion on one piece of these much larger problems. Hopefully, it is a long way from the end.