The quantitative study of security dynamics in computer-science technologies is essential for understanding security-development patterns of information systems. Here, we specify and investigate security dynamics as (i) the relation between technological and security developments, (ii) the security development, modelled as the evolution of security considerations among technologies, and (iii) the effect of security development on the opinion given to technologies. We perform a scientometric analysis on arXiv e-prints (n=340569) related to 20 computer-science technology categories. Our empirical results are threefold. First, we provide evidence of a lack of relation between the technological and security developments: while most categories follow a sigmoid-growth curve of technological development, this latter is not a determinant of security development. Second, we find a security-attention pattern: over the lifetime of categories, security considerations appear more frequently, emphasising that security gains more attention at a later stage of technological development. Third, we find an opinion pattern: the experts' opinion related to each category is positively explained by the prevalence of security considerations. These results emphasise new methods for understanding, modelling, and benchmarking security dynamics of technologies, which brings new heuristics for considering changes related to the security of information systems.