Image-based Communication on Social Coding Platforms (2306.15851v1)

Abstract: Visual content in the form of images and videos has taken over general-purpose social networks in a variety of ways, streamlining and enriching online communications. We are interested to understand if and to what extent the use of images is popular and helpful in social coding platforms. We mined nine years of data from two popular software developers' platforms: the Mozilla issue tracking system, i.e., Bugzilla, and the most well-known platform for developers' Q/A, i.e., Stack Overflow. We further triangulated and extended our mining results by performing a survey with 168 software developers. We observed that, between 2013 and 2022, the number of posts containing image data on Bugzilla and Stack Overflow doubled. Furthermore, we found that sharing images makes other developers engage more and faster with the content. In the majority of cases in which an image is included in a developer's post, the information in that image is complementary to the text provided. Finally, our results showed that when an image is shared, understanding the content without the information in the image is unlikely for 86.9\% of the cases. Based on these observations, we discuss the importance of considering visual content when analyzing developers and designing automation tools.