The Cultural Impact of Content Management Systems: A Topic Modeling Approach
Website development and content creation has transitioned from, in the 1990s, a largely manual process controlled by webmasters or other technical staff to a process that is assisted by content management systems (CMSs) accessible to non-technical individuals (Souer, Urlings, Helms, & Brinkkemper, 2011). Among sites built and maintained using CMSs, it is estimated that around 50% use Wordpress, a popular, open-source system originally developed for blogging (“CMS technologies Web Usage Statistics,” 2015). The first claim this presentation makes is that CMSs are a promising research area that has received little attention, especially from the humanities. Other than a few papers taking an engineering perspective on CMS design (e.g., Aggarwal, Prakash, & Sofat, 2009; Souer & Joor, 2010) or reports on implementation (e.g., Alejandro Garza, 2009), there has been little work on CMSs and none on Wordpress specifically. However, as noted above, the world is increasingly represented through these systems. As Dourish (2014) notes in relation to database standards, the practice of organizing and presenting data and conceptions of the world are mutually influencing. CMSs establish content ontologies (e.g., posts, pages and image galleries) as well as taxonomies for how content is organized (e.g., categories and posts). A system as common as Wordpress, then, has considerable influence on how the world is represented and warrants attention as both a technical system and a cultural form. This connection between technical details and culture is the basis for much work in platform studies. However, while work in that area (e.g., Montfort, 2009) has tended to focus on gaming systems and other overtly creative systems, it has not yet paid significant attention to platforms that support work and other forms of representation such as blogging or running an online store. Work in software studies (e.g, Manovich, 2013) has, to a lesser extent, exhibited similar biases. From this perspective, Wordpress represents an opportunity to better understand how technical systems influence the kinds of knowledge work that Liu (2004) describes as increasingly pertinent to both business and humanities interests. My second claim is methodological. As trends in software studies (e.g., Gehl & Bell, 2012) show, history is important to understanding how systems and applications that often seem natural are in fact the products of decisions and negotiations. However, large software systems are difficult to grasp with traditional humanities methods. Drawing on work in computer science that uses topic modeling to monitor software development (e.g., Kuhn, Ducasse, & Gîrba, 2007; Thomas, Adams, Hassan, & Blostein, 2011), I will present data (fig. 1) on the historical development of Wordpress. Topic modeling 100 release versions of Wordpress gives initial access to what Dourish (2014) refers to as “the technological fabric of information processing that shape[s] the emergence of particular forms of both technology and technological practice.” This presentation of initial results will: 1) discuss the use of topic modeling for understanding software history from a humanities perspective 2) address specific considerations related to visualization in this context and 3) identify trends specific to Wordpress that warrant further work.