Open Code Research From A Multidimensional Perspective

In this study,open code refers to the process of freely distributing code on some public host platforms;therefore,other users are free to get the code,run the code to verify the results and analyze the code themselves.Open code is becoming a prominent way of scholarly communication,affecting scientific dialogue,thoughts,and behaviors.First,open code enables researchers to improve the discoverability and visibility of their work through rapid access to code resources;secondly,open code is a new requirement for scientific and technological information exchange,which makes it easier for researchers to connect with each other by creating new opportunities to interact with and contribute to ongoing public projects;thirdly,reproducibility is regarded as the baseline of science,and open code is the prerequisite for achieving the reproducibility of scientific research.Publishing source code and its version history is essential to ensure the quality,reliability and reproducibility of scientific research.This study provides an in-depth understanding of open code from a multidimensional perspective,including social environment factors,journal policy factors,researchers' behavior factors and the influences of open code.In order to promote open code work in China,the author designs policy model to better promote China's open code action,promote the open movement,expand the scholarly communication mode,and provide support for the reproducibility of scientific research.The social environment factors of open code include international political factors,social cognitive factors and technological development factors.The introduction of policies provides guarantee for open code development,the changes in cognitive and practical environment provide the impetus for open code development,technology innovation supports the open code development.Research on journal policy factors of open code mainly analyze the relevant characteristics of the open code policies in biomedical journals,and the impact of open code policies on the open code behavior of scholars.On the whole,most journals in the biomedical field have different levels ofrequirements for open code,but most journals do not explicitly mention how to deal with open code.The open code policies of open access journals are more common than traditional subscription journals.In general,public online repositories are the platforms for publicly available code recommended by most journals,and high-impact journals are more likely to open code than low-impact journals.By analyzing the editorials of Nature Methods relateds to open code over the years,the attitudes of the journal towards open code is derived,and then the impact of open code policies on author behaviors is analyzed.The number and percentage of open code papers in the post-policy period were much higher than in the pre-policy period,and the number and percentage of open code papers experienced a sharp rise from the pre-policy period to the beginning of the implementation of the open code policy.The number and percentage of open code papers in academic papers published in the pre-policy period showed a downward trend.After the implementation of the open code policy,the number and percentage of open code papers in the post-policy period showed an upward trend.Research on researchers' behavior factors of open code select the conference papers of the top conferences in the field of computational linguistics as research samples,and analyze the characteristics of the open code behaviors of the authors by analyzing the location,storage location and accessibility of the open code in the papers.The analysis finds that open code papers are not high in the top conferences of computational linguistics,but the overall trend is on the rise;the open code is mainly concentrated on pages 1-2and 9 of the papers,lines 0-400(especially lines 0-200),and the chapters of introduction,conclusion,and experiment;starting in 2011,public platforms have gradually become the choices of authors to host open code;after 2012,the conditions of accessibilities of open code are good.Research on the influence of open code has selected the citation counts and altmetrics tools as measurement indicators.The citation impact of open code is reflected in: In most years,open code papers have more citations than papers without open code.Open code can be a positive factor in promoting short-and long-term citations.In the small citation window,intra-domain citations are more susceptible to open code.However,when it comes to big citation windows,inter-domain citations are more susceptible to open code.Open code accessibility significantly affects the citation impact of the papers.Code on both public and private platforms has a significant positive impact on citation counts,but the code hosted on the public platform has a stronger impact.Open code has an impact on the Altmetric Attention Score of the papers.However,among the Altmetrics measurement indicators,open code of academic papers is related to the Twitter mentions and is not relevant to News mentions and Blog mentions.The averages and maximums of News mentions and Blog mentions of the papers that provide code downloads are higher than those of the papers that do not provide code downloads,and the code download of academic papers affects the attention of the papers on blogs and Twitter.The way how code opens affects Twitter mentions.Finally,the author condenses the policy model that promotes open code work in China.The model includes policy model design goals,principles,methods,and policy metamodel construction,and creates different policy tools.Specifically,it includes the introduction of open code related policies and regulations,the development and promotion of code reference national standards,encouraging journals to choose the appropriate code publishing model,improvement of research funding mechanism,incorporating open code into the research evaluation system,building open code guidance mechanism,etc.