PREreview of "Understanding Main Path Analysis"
This Zenodo record is a permanently preserved version of a PREreview. You can view the complete PREreview at https://prereview.org/reviews/21608208. This paper aims to offer a deeper, more rigorous understanding of main path analysis, a method used relatively frequently in the study of citation networks. The paper also introduces a new approach to main path analysis referred to as search path entropy. This is a variant of the traditional search path count approach. The paper is rather long and complex. Because of time constraints and my limited expertise on some of the mathematical aspects of the paper, I must admit that I have not been able to study the paper in full detail. This review report is based on a careful high-level reading of the paper. I consider the paper to be of high quality. Although there are some aspects of the paper that I have not been able to study in full detail, the research reported in the paper seems rigorous to me. I very much agree with the authors that main path analysis lacks a solid theoretical foundation. Given the relatively frequent use of the method, this is a significant problem. The paper offers valuable insights into the properties of main path analysis. In addition, the search path entropy approach introduced by the authors appears interesting. It seems to offer an attractive alternative to the conventional search path count approach. I also agree with the argument made by the authors in favor of a 'basket' approach over a traditional approach focused on identifying a single 'main path'. I would like to use this opportunity to draw attention to my own closely related work, which the authors may not be familiar with. Together with colleagues, I developed an alternative to main path analysis based on the concept of intermediacy. We published this in the following paper: https://doi.org/10.1098/rsos.190207. In our paper, we conclude that compared to main path analysis "intermediacy offers a more principled approach for tracing the historical development of scientific knowledge". Like the authors of the paper under review, we go beyond the identification of a single main path and we instead allow for the identification of multiple important paths. However, there is a major difference between the paper under review and my own paper. The paper under review recommends the longest path between two nodes as a practical alternative to the path identified by main path analysis. My own paper more or less points in the opposite direction. Based on theoretical arguments, it essentially argues in favor of shorter paths over longer ones. The two papers appear to start from different theoretical ideas and therefore arrive at conclusions that seem to partly contradict each other. This might warrant further analysis and discussion. Finally, I would like to point out some minor inaccuracies in the paper: Eq. 2.8a: It seems the second occurrence of G^(spe) should be G^(spc). "This subgraph of the infinite lattice.": The sentence is incomplete. "using eitrher of the SPC or SPE weight baskets": Typo. "so real-worl networks have no well-defined geodesic": Typo. "the just the nodes of zero unit criticality": Typo "contains the nodes lying a main path": The word 'on' seems to be missing. Competing interests H.C.W. Price, the first author of the paper under review, and I are both involved in the Meta-Moment project currently carried out by the Research on Research Institute (RoRI). Competing interests H.C.W. Price, the first author of the paper under review, and I are both involved in the Meta-Moment project currently carried out by the Research on Research Institute (RoRI). Use of Artificial Intelligence (AI) The author declares that they did not use generative AI to come up with new ideas for their review.