I arrived in Portland with a plan to get started on Microsub during the event. Recruitment was conducted with a plan to stop scheduling interviews once their content became repetitive-i.e. Interviews were conducted with 15 IndieWeb contributors. Finally, Chapter 7 uses trace ethnography (Geiger and Ribes 2011) and interviews to investigate how IndieWeb’s syndication relationship with the "corporate web" influences development and maintenance. 2011) used Wikipedia as a large corpus to train a topic model for analyzing sparse documents. One method of addressing this challenge is to train a topic model from an external corpus. Network data about GitHub contributions was linked with network data and an LDA model about IndieWeb’s chat archives. People posting to IndieWeb’s chat sometimes prepend or append their usernames with contextual information. 2003) to identify topics of discussion across IndieWeb’s chat archives. And even though contributors to IndieWeb’s chat, wiki, GitHub repositories and other resources have implicitly accepted that their data may be accessed by anyone, this does not mean they automatically consent to being quoted in research publications. At the beginning of the interview, participants were asked how they would like to be identified in this dissertation and related publications.



The most important takeaway from boyd’s speech is that the field of technology, like many others, cannot move forward without reckoning with injustices in which it has been complicit. However, for the purpose of addressing my research questions, focusing on cultural diversity to that extent could misrepresent the reality of working on projects like IndieWeb. However, as I progressed in this dissertation, I realized that IndieMap was not as well suited to addressing my research questions as the data I have analyzed from GitHub and IndieWeb’s chat archives. In this study, my purpose is to generally identify subjects of conversation as a means to evaluate the extent to which individuals are active in conversations about IndieWeb’s principles as well as other broad-stroke topics such as events and event-planning, online community maintenance, and IndieWeb-specific technical objects. Although individual messages are often short and messy, conversations are typically held over multiple messages in a short time period. One challenge posed by applying LDA to IndieWeb’s IRC logs is that individual messages are short and meaning may be conveyed across multiple sequential messages. Nonetheless, as described in previous studies of IRC (Mutton 2004; Tuulos and Tirri 2004), recipients of each message can be inferred using heuristics based on explicit references to an individual’s chat nickname or the time between messages.



After this cleaning, the number of messages was reduced from 1,159,559 to 923,634. These steps prepared the chat logs for further analysis, which will be described soon. However, IndieWeb’s wiki contains thorough accounts of a large number of projects. This could consist of direct contributions to software or standards, substantial customizing of their own IndieWeb site (defined as a site that used Microformats and Webmentions), or ongoing participation in IndieWeb’s chat discussions or wiki. Documents range in length widely, from a 29-page collection of notes about one month’s worth of software development, to less than a page of commentary on a day’s chat logs. This was the first argument I encountered from one of my interview participants that IndieWeb’s barrier to entry could be a virtue unto itself. Most specifically, I adopted Microsub largely because it was an emerging norm, something I believed other community members would take up, and thus a good tool through which to understood IndieWeb’s community. Therefore, I took the notion of core/periphery as a guideline to ensure I interviewed people who reflected varying positions among IndieWeb’s community. Further, Phalogenics Male review although those who considered themselves to be less technologically adept may have felt less influential, there was a degree to which they could still be influential by identifying challenges that others had not experienced and otherwise broadening the diversity of needs that were acknowledged by IndieWeb’s community.



Technical expertise was commonly cited as a factor when identifying influence from others. Studies of online communities have often incorporated conventional ethnographic methods alongside novel forms of analysis suited to the nature of digital data. In this section, I describe how I have collected logged data for analysis. By default, GitHub repositories are public, and their data may be accessed either via a Web browser or through GitHub’s API. However, they have been classified separately from "IndieWeb building blocks" because a significant portion of the documents reviewed are related to technologies developed outside of IndieWeb’s community, such as servers, content management systems, and Web frameworks and languages. One possibility would be to use the text from IndieWeb’s wiki, which includes similar language as its chat but is more structured and consists of longer documents. One individual wrote in chat that it could be awkward navigating IndieWeb source code because it is spread across individuals’ largely unpolished GitHub repositories. I also identified how concepts established during coding related with one another to describe tensions among goals, activities, and various structures, and approaches for resolving those tensions. This was appropriate for the exploratory nature of this analysis, in which I sought to identify if multiple variables could converge, defining ‘roles’ based on combinations of topics used together, proportions of different GitHub activities, and other variables.