The innovation dynamics of programming technologies

Programming technologies evolve rapidly. Yet, the factors related to the rise and fall of programming technologies have not yet been revealed. To close this gap, we study innovation in programming by analysing data from the online coding platform Stack Overflow. Our aim is to understand how competition affects the growth trajectories of technology tags over time. Using correlation networks that encode dynamic tag usage patterns, we identify two robust technology clusters. They represent (i) core computing facilities covering operating systems, databases and servers, and (ii) application development technologies, containing frameworks for web development and machine learning. We find that declining old technologies are primarily associated with the core computing facilities cluster, while rising new technologies are mainly associated with the cluster of application development technologies. We derive common factors associated with the rise and fall of technology tags on the platform: technologies that link positively to other new technologies and negatively to any frequently used, old technology have higher chances of gaining traction and becoming successful. We conclude that popular, rising technologies tend to supplement rather than complement existing technologies. The empirical findings point towards creative destruction as a mechanism that shapes the innovation dynamics of programming technologies.