1. Introduction
[1.1] In 2019, the social media platform Tumblr employed hundreds of software engineers, who were paid on average $123,882 a year (Glassdoor 2020). However, the Tumblr community's most beloved programmer, who created a program that many Tumblr users relied on every time they logged in, was a lone, unpaid volunteer developer, a mysterious, benevolent figure often just known as "the Xkit Guy" (Romano 2013). From 2011 to 2014, the Xkit Guy single-handedly maintained the Xkit Browser Extension—a piece of code that Tumblr users could use to change how the website functioned on their own computers. Unlike the official Tumblr developer team, who are often seen as out of touch with the needs and desires of the fannish communities that make Tumblr their home, the Xkit Guy was a member of these communities and created code that served these fannish users' needs.
[1.2] The Xkit Browser Extension is a novel example of fans using unlicensed code as a form of opposition to the owners of the social media platform on which these communities make their home. Previous scholarship of fan resistance has surveyed a variety of strategies (fan art and fan vids, fandom meta, and political and economic action) that fans have employed to oppose the rights holders of their object of fandom (Freund 2018; Jenkins 2006; Seymour 2018). Scholars have also previously chronicled a variety of subversive practices that fans have used to escape censorship or manipulate algorithms on the platforms they inhabit (Andrejevic 2009; Jenkins 2006; Stanfill 2019). However, Tumblr provides a strange case of a social media platform where fans not only organize communities on the platform but also take up fannish attitudes toward the platform itself (i.e., "the Tumblr Fandom"). Similarly, these users create fan works that actually alter the platform itself—just as fans of a TV show might create fix-it fan fiction to solve perceived problems with their object of fandom, Tumblr users created the Xkit Browser Extension as a piece of code to fix Tumblr.
[1.3] To make sense of this sort of unconventional fan work, I draw on the existing literature of resistant fan works (Tushnet and Coppa 2017; Jenkins 2006; Goodman 2015), game studies literature on video game modding, (Taylor 2006; Postigo 2010; Kücklich 2005; Sotamaa 2007; Scacchi 2010), and Lawrence Lessig's formulation of "code as law" to theorize social media modding as a transformative work in the medium of "outlaw code" (Lessig 1999; De Zwart and Humphreys 2014). Using this framework, I situate the Xkit Guy in the context of previous debates about fan resistance, fan labor, and fan exploitation, and chronicle how the volunteer labor of fannish software developers empowers the Tumblr community against the profit motives of the corporations that control the platform.
2. Welcome to the Blue Hellsite
[2.1] In order to understand the Xkit Guy, we must first understand the ecosystem he existed in: Tumblr, from 2011 to 2018. Tumblr has been described by its founder David Karp as a "microblogging platform" (Al-Muslim 2017), but it is much more like a social media site. Though the platform allows users to create individual blogs, most users interface with content on the platform through a dashboard (or "dash" for short), a feed of posts from the blogs they follow. Users can like posts, reblog posts to circulate them, or reply to posts by adding tags or comments, much like on other social media platforms. However, Tumblr stands out among mainstream social media platforms as a special hub of fan communities as well as a variety of other subcultures. These communities were initially drawn to Tumblr by peculiar quirks of the platform's design and history and its original affordances for privacy and lack of censorship. This history has forever shaped Tumblr's demographic makeup as well as the relationship between the owners of the platform and the communities that make the platform their home (McCracken 2017; Morimoto and Stein 2018).
[2.2] Tumblr is widely acknowledged by scholars as a major site of fan activity. As fans began to migrate away from LiveJournal, Fanfiction.net, and Delicious in 2009 to 2012, many of them found a home on Tumblr (Bury 2017; De Kosnik 2016; Kohnen 2018; Stein 2018). Tumblr has traditionally been a point of interface between the academy and fandom, a site where acafans can conduct research and share their findings with the community on blogs like Fansplaining, Fanalytics, and the Learned Fangirl (Morimoto and Stein 2018; Downey et al. 2018). In fact, in 2018, Transformative Works and Cultures ran a special issue focused exclusively on Tumblr (Morimoto and Stein 2018). Tumblr is also famous for creating a space for pollination and migration between fandoms, creating hybrid fandoms such as the infamous SuperWhoLock.
[2.3] Tumblr is also home to a variety of artistic production communities (including animators, comic artists, and amateur erotica producers), and fringe subcultures and affinity groups—including queer advocacy groups and disability rights communities (Petersen 2014; Cho 2018; Fink and Miller 2014; Schott and Langan 2015; Tiidenberg 2016). Though not all of these groups are explicitly fannish, many of them have the same requirements from a platform that fan communities do, and they flourish in similar spaces. Furthermore, these other communities inhabiting Tumblr are deeply entwined with Tumblr fandom—Alyssa Smith (2017) and Lori Morimoto and Louisa Stein (2018), for instance, have all written on how Tumblr fans existing in a milieu of queer subculture and disability activism led to the formation of "the Discourse," the particular style and rhetoric of identity politics that characterizes many conversations in Tumblr fandom and distinguishes it from fandom on other sites.
[2.4] I will provide a brief history of how the affordances of Tumblr made it a nourishing home for this multitude of subcultures, drawing on histories of Tumblr from the queer studies, disability studies, and platform studies literature. I draw from these other literatures both to demonstrate that these authors' findings can also be applied to fandom and to suggest that my own findings about fandom describe a more general set of dynamics between an internet subculture and the platform they organize on.
[2.5] In Tumblr's early years, many subcultural groups found the platform welcoming because it did not censor certain forms of content and, in general, did not share many social media platforms' optimized-for-eyeballs-on-ads architecture. As internet studies scholars such as Tarleton Gillespie (2018) have noted, most mainstream social media platforms (such as Twitter, Facebook, and Instagram) profit from mining their users' data to sell targeted advertising, which in turn encourages these platforms to adopt a set of design features that serve this profit model. These platforms require users to maintain a single, persistent identity corresponding to their real-world selves, which can be tracked and targeted with ads (Gillespie 2018; Fuchs 2014). They use algorithms to prioritize or deprioritize certain content in order to maximize engagement. They are heavily moderated—in terms of sexually explicit or copyrighted content—to eliminate content that might expose them to liability or that advertisers might find distasteful (Andrejevic 2009; Gillespie 2018; Cho 2018).
[2.6] However, when Tumblr was created in 2008 it had no ads, no data tracking, no requirements for a persistent identity, and virtually no moderation or algorithmic curation at all (Gillespie 2018; Morimoto and Stein 2018). This combination of affordances for privacy, anonymity, and uncensored self-expression created a nurturing environment for the flourishing of subcultures and fandom (Petersen 2014; Cho 2018; Morimoto 2018). As Morimoto and Stein (2018) pointed out, "Tumblr's interface specificity—its affordances and limitations—have shaped fandom uses of the site and arguably have thus shaped fandoms themselves" (¶ 3.3).
[2.7] Tumblr was a comfortable space for subcultures that value anonymity because it did not require users to maintain a single persistent identity or to disclose information about their real-world selves. Unlike Facebook, which required a "real name" (Gillespie 2018), all Tumblr required to sign up for an account was an email address. And whereas other social media platforms tried to restrict users to one account, Tumblr made it easy to create multiple disconnected side blogs for one account—meaning Tumblr users could separate their fannish or subcultural activities from their main public persona (Tiidenberg 2016). Tumblr's faulty search features also made it difficult to seek out specific posts, adding a level of anonymity by obfuscation (Cho 2018; Haimson et al. 2019).
[2.8] Though Tumblr's content is technically a persistent public archive, the impossibility of searching it practically makes that content private and ephemeral. Increased anonymity made Tumblr a safe space for members of fringe or taboo subcultures who might not want their posting to be discoverable. For instance, Tumblr became a haven for closeted queer teens to explore their sexual identities in a sort of private public (Cho 2018; Fink and Miller 2014; Haimson et al. 2019). These same concerns made Tumblr appealing to a variety of other folks who might not want to be outed about their online activities—such as amateur porn producers (Tiidenberg 2016) or mental health support communities (Schott and Langan 2015). And because fannish activities can often be a source of shame and many fans do not want their fannish activities to go public (Bacon-Smith 1992), Tumblr's default anonymity was similarly welcoming to fans (McCracken 2017).
[2.9] Tumblr's early lack of censorship also helped a variety of subcultures, including fandom, flourish. Up through 2018, Tumblr was one of the only social media platforms to not censor nudity and pornography (Tiidenberg 2016; Cho 2018). On an internet where explicit content is banned on most forums, Tumblr was a boon—a polished and reputable site that would not ban erotic content outright. Kink and erotica communities found a home on Tumblr, as did political movements that circulated nude photography—for instance, the body positive or sex positive movements (Tiidenberg 2016). And, as Alexander Cho (2018) and others have argued, lack of censorship was also important to gay and trans users who did not post sexually explicit content but might have their posts discussing queerness flagged as "inappropriate for children" on other, more heavily moderated platforms (Fink and Miller 2014; Haimson et al. 2019). Though fan production is not all sexual in nature, the possibility for unbridled self-expression and posting erotic works has historically been important to the flourishing of fandoms, which often become socially organized around, for instance, writing slash fic or kink-memes (Bacon-Smith 1992; Russo 2017). Seeking an outlet for sexual expression has been historically understood as a major driver that brings fans, especially female fans, to fandom (Penley 1992; Bacon-Smith 1992). Though text-based erotic fan fiction has had an easier time escaping censorship, fans also create erotic fan art and fan vids (Dennis 2010; Freund 2018), and fans who chose these mediums to express themselves needed Tumblr's affordances. Furthermore, whether or not it directly impacted the circulation of text-based erotica, the lack of censorship on Tumblr signaled that it was a welcoming space where no one else got to set the rules for how fans were allowed to present themselves.
[2.10] This history of Tumblr also informs my methodology. I will draw on the archive of publicly available Tumblr posts to construct a history of events in the Tumblr community; however, to protect and perpetuate the affordances—anonymity, privacy, and ephemerality—that make Tumblr a sustaining home for the communities I write about, I have adopted a citation practice designed to not deanonymize or draw undue attention to the users I will reference. When I refer to obviously public-facing Tumblr blogs, such as the Xkit Developer Blog, I will provide direct links to the posts I cite. But in cases where I cite the personal blog of a private individual, I have omitted a hyperlink or blog title to protect the privacy of the user. Posts cited this way should be read not as specific references but as representative examples that illustrate a general trend.
[2.11] Tumblr in 2020 is a very different place from the site I have described here. Even in 2013, when fandom had only recently become established on Tumblr, a number of these functionalities were already starting to change. Tumblr had been bought by Yahoo (and it since has changed hands to Verizon as part of their Yahoo purchase, and then to Automattic), which was beginning to profit from advertisements and was implementing some algorithmic reshufflings of the platform. Many of the communities I have described no longer have a home on Tumblr—most users point to December 17, 2018, the day that Tumblr began using algorithmic curation to ban explicit content, as the final nail in the coffin of the old Tumblr on which these subcultures once thrived. Though the initial architecture of Tumblr created ideal conditions for a variety of fannish and subcultural communities, the optimal conditions for these communities' flourishing have always been at odds with Tumblr's profit motives as a platform. In the following section, I discuss how the struggle between the communities that inhabit Tumblr and the platform's owners fit into a general pattern of fraught relationships between fans and the platforms they organize on—as well as some ways in which Tumblr users' relationship to the platform and its staff must be considered unique.
3. Fans and platforms: Labor, exploitation, resistance
[3.1] In order for online fandom to flourish, fans need online spaces to organize in. On an internet where, as platform studies scholars tell us, the digital commons are increasingly becoming privatized, these spaces are usually owned by somebody else (Gillespie 2018; Scholz 2016). Fan communities have on occasion created volunteer-run, community-owned spaces for hosting their interactions and artworks, such as the fan fiction repository Archive of our Own (AO3) (De Kosnik 2016). However, interventions like this are rare, especially as individually owned small forums and websites are increasingly crowded out by large, corporate-owned platforms as part of the process of platformization of the Web 2.0 internet (Jenkins 2013; Scholz 2013; Andrejevic 2009). The majority of fan activities on the internet, from distributing fan art to discussing media, take place on corporate platforms such as Tumblr, Reddit, DeviantArt, and Twitter. Because the profit motives of the owners of these platforms are usually not fully aligned with fan communities' needs and desires, fandoms (and other subcultures) frequently find themselves in conflict with the landlords of the spaces where their communities live.
[3.2] Social media platforms operate by exploiting the creative production (fannish and otherwise) of the communities that take up residence on them. As Tiziana Terranova (2000), Trebor Scholz (2016), and others have argued, the enclosure of the digital commons by platforms is a process by which platform owners claim ownership and profit from labor that is freely given in internet communities. Platform users, including fan communities, might bring eyeballs and advertisers to the platform, but the platform owners, not the users, get to collect the profits (Jenkins 2013; Fuchs 2014; Banks and Humphreys 2008). Fan works often essentially constitute free advertising for both media producers and platform owners (Stanfill and Condis 2018; Stanfill 2019).
[3.3] By posting on platforms, users, including fan producers, often give up their intellectual property (IP) rights to the work they post (Andrejevic 2009; Fuchs 2014). Platform owners are free to police content and make arbitrary decisions—what content should be taken down under the Digital Millennium Copyright Act (DMCA), what violates the community guidelines—which allow them, to some degree, to force fan production to stay within the lines (Andrejevic 2009; Jenkins 2013). Platforms will leap much more quickly to protect a media corporation's IP rights than those of an independent artist or fan (Andrejevic 2009). Essentially, fans are forced to pay the products of their own creative output and their community's cultural capital as rent on the platforms they organize on.
[3.4] Fan communities on platforms also depend on affordances of platforms that they have no control over, putting them almost completely at the mercy of platform developers. As we have seen with Tumblr, small design decisions by platform owners can have major effects on how fan communities are organized. For instance, many authors attribute the rise of fan art, gif sets, and other visual mediums in fandom to the simple fact that Tumblr hosted these sorts of media and made them easy to distribute (Morimoto and Stein 2018). Fannish communities can be extremely vulnerable to change in platform affordances. For example, in 2011, the vibrant fannish community organized on the bookmarking site Delicious all but vanished overnight when the site's developers made a change that eliminated searches for tags containing punctuation—including, disastrously, the slash character (Cegłowski 2013). Writing outside of fan studies, Phillip Bagust (2014) has likewise written on how the changes that Yahoo made after purchasing Flickr in 2013 to make the platform more commercializable—a process he calls "Facebookification"—destroyed the communities that had sprung up on the site. Fan communities as well as other online subcultures have little to no say in how the spaces where they live are governed, and a few keystrokes by a platform developer can spell disaster for a community.
[3.5] Fans struggling over properties that they do not hold the deeds to is not a new phenomenon in fan studies. As Henry Jenkins noted in Textual Poachers (1992), most fans do not own the IP rights to the media that they love—an arrangement that can be distressing for fans, especially when they feel that the current author or showrunner is screwing it up. Fans deal with their feelings of moral ownership over these media properties that they do not legally own by producing their own fan works in order to, Robin Hood–like, reclaim these commons for their communities and "protect [them] against abuse from those who created [them] and have claimed ownership over [them]" (Jenkins 2006, 59). As Catherine Driscoll (2006) writes, fans' relationship with their objects of fandom is often a mixture of "fascination and frustration"—fascination because they are attached to these texts, but frustration because these texts are written by others without their input, and thus will "never fully conform to audience desires" (86). And so fans are driven to create fan works that fulfill these desires themselves.
[3.6] Early fan scholars were excited by women who used fan fiction to satisfy their narrative and sexual desires in ways that the TV shows they watched would not (Bacon-Smith 1992; Penley 1992; De Kosnik 2009). When queer fans do not see representation of different genders and sexualities in the media they watch, they write them in with slash fic and fan vids (Tushnet and Coppa 2017; Russo 2017; Lothian, Busse, and Reid 2007). When fans of color do not see Black, Latinx, or Asian people onscreen, they create them through racebending fan works and cosplay (Seymour 2018; Pande and Moitra 2017; Lopez 2012). When fans feel that the official rights holders have mismanaged their object of fandom and refused to fix problems with it, fandoms have traditionally had a solution—make their own unsanctioned, transformative fan works to fix it (obsession_inc 2009).
[3.7] Fandom also has a history of pushing back against attempts—including attempts by platform owners—to exploit their communities. Since fan studies' origins in subculture studies, fandom has been understood as a resistant and countercultural kind of community that opposes hegemonic control of the media and commercialization (Hall 1971; Hebdige 1979). Both Henry Jenkins (2013) and Mark Andrejevic (2009) have written on how fan production does not always allow for easy exploitation. Unruly fan creators cannot always be wrangled into producing content that is easy to monetize—they make things that are unpalatable to advertisers, or inscrutable to corporate executives (Andrejevic 2009). Though fans are constrained by the affordances and algorithmic laws of a platform, they can still organize themselves in ways counter to a platform's profit model.
[3.8] The Xkit Browser Extension merges these two forms of fannish resistance: it is a transformative fan work (in the medium of code) that opposes corporate platform ownership and monetization of Tumblr communities. But in order to understand Xkit as a fan work, we first need to understand the Tumblr community as a fandom, which is already participating in more traditional forms of fannish resistance.
4. "Thanks, I stole them from the president": The Tumblr fandom
[4.1] Even among social media users, Tumblr users are a highly organized community, so much so that we might view them as a fan community in their own right. As Amanda Brennan writes, "When I meet Tumblr users, you talk about Tumblr and you feel like you're part of something much bigger. You understand the language, you speak the memes, and it's like a code word [sic]" (quoted in Morimoto 2018, ¶ 37). Whereas Twitter or Facebook users may not see the whole site as their public sphere, Tumblr users see themselves as part of a single imagined community. Inside references and code phrases such as "I like your shoelaces" (the coded public greeting to indicate that you recognize another person as a Tumblr user) help Tumblr users solidify group identity, similar to how members of other fandoms might use coded references to their favorite show or book.
[4.2] Fannishness as a way of relating permeates Tumblr. Tumblr users tend to use the term "fandom" liberally to talk about all groups that organize on the platform—"the Geology Fandom," "the Ancient Greece Fandom"—or even to describe the entire membership of the platform—"the Tumblr Fandom." In 2014, there was even a fan convention devoted to Tumblr—the infamous Dashcon. And there is already precedent for fans appropriating an entire platform as an object of fandom in this way—for instance, Maciej Cegłowski (2013) of the social bookmarking site Pinboard noted that in 2011 fans on Archive Of Our Own were writing Delicious/Pinboard slash fic. For these reasons, it is helpful to think of Tumblr users not just as a conglomeration of fans of other media but also as a fandom of the social media site itself.
[4.3] Like many fandoms, Tumblr users are a highly introspective community who are aware of how the site's affordances affect their way of life. Tumblr users frequently discuss how Tumblr's structure influences their communities—a practice we might regard as adjacent to writing fandom meta. For instance, users have been documented arguing that Tumblr's reblogging structure is responsible for debates on the platform getting out of hand (Morimoto and Stein 2018), or attributing their comfort expressing their queer identity to Tumblr's anonymity (Cho 2018). Tumblr fans are highly protective of the features and affordances that their communities depend on—for instance, "bring back replies!" (a call for the platform to reinstate a feature that let you reply privately to posts) is a common refrain among users (Morimoto and Stein 2018). After each new update to the site, Tumblr is flooded with posts discussing how the new features work and whether they help or harm Tumblr communities.
[4.4] Tumblr users are also very aware of the power that the Tumblr administrators have to negatively alter their object of fandom and destroy existing communities. Tumblr fans are highly suspicious of the Tumblr administrators, often referred to as "The Staff" because of the staff.tumblr.com blog that provides the Tumblr community's main point of contact with the developers. Users despise The Staff, both because they see them as incompetent and because they fear that The Staff are motivated by profit interests and do not have their community's best interests at heart. When Tumblr was bought by Yahoo in 2013, for instance, there was a general uproar, and users expressed fears that new management would change Tumblr for the worse. The Tumblr community's relationship with The Staff is not unlike an antagonistic relationship between a fan community and a hated showrunner—in fact, Tumblr fans talk about The Staff messing up Tumblr in very much the same way they would talk about Steven Moffat messing up the latest season of Sherlock. It is not uncommon to see users post comments to the tune of "I'm gonna burn down tumblr hq I stg [swear to God]."
[4.5] Tumblr users revile the idea that the Tumblr Staff are profiting off their fandoms, and they celebrate whenever the platform proves itself un-monetizable. For instance, when Yahoo was forced to sell the platform, which they had bought for $1.1 billion, to Automattic for less than $3 million, fans cheered and took credit (figures 1 and 2). Users gleefully circulated memes celebrating the "tumblr gremlin userbase," who had resisted being harnessed as eyeballs-for-cash: "hey @yahoo was banning porn really worth losing 1 billion dollars?" or "through sheer spite…the userbase DECIMATED the value of a website." Like the YouTube users described by Andrejevic (2009), Tumblr users are unruly creators who resist being harnessed for profit—and celebrate their unruliness.
Figures 1 and 2. An example of the memes and gifs Tumblr users circulate celebrating Yahoo and Verizon's inability to profit off of their communities.
[4.6] Rather than just lying down and accepting software and policy changes that hurt their communities, Tumblr users organize to resist unwanted changes to their platform. To present one case study, in 2017, The Staff rolled out the "Best Stuff First" feature, which was automatically met with huge pushback. This on-by-default setting algorithmically moved high-engagement posts to the top of a user's dash. Tumblr users feared that this feature, which prioritized already popular posts, would hurt struggling or emerging artists and fan creators. In response, Tumblr users circulated posts alerting other users to the feature, and instructing them on how to turn it off. For example, one prominent Tumblr artist used their platform to write, "Support ALL content creators and turn [best stuff first] off, please. It sucks having our work be pushed to the bottom never to be seen." Other users voiced their negative experiences: "Many people that we reblog here are likely affected by this terrible change and have noticed a drop in notes because tumblr is suddenly hiding their content and pushing names that are already big or are paying tumblr to be advertised." These guerrilla information campaigns used Tumblr's own features for circulating posts to help minimize the negative effects of a new feature on their community. In this example, we see how Tumblr users feel entitled to having a say in what features get added to the platform and will take coordinated action to resist unwanted changes.
5. Enter the Xkit Guy
[5.1] Like the Tumblr users previously described, the Xkit Guy and Xkit users are fannish rebels who push back against the corporate power of the Tumblr platform in order to protect affordances and functionalities that their communities rely on. But unlike these other cases of fan resistance, the medium of the Xkit Guy is code. Rather than circulating resistance on the platform, the Xkit Guy—and users of his software—change the platform itself. Although Xkit and other browser extensions cannot directly modify Tumblr's code base (a realm inaccessible to users), they can modify how Tumblr's interface operates on a user's computer (a realm in which the user has total control). While Xkit is effectively a skin that runs on top of Tumblr's actual code, its modifications to how Tumblr is displayed can have deep effects on the site's usability, preserving features users find essential and allowing new possibilities for fannish organization.
[5.2] "The Xkit Guy" is actually a mythical title, which has been held collectively by a set of developers who have at various times taken up the role of maintaining the Tumblr Xkit Browser Extension. These volunteer developers have taken on a folkloric, almost messianic image in the mind of the Tumblr community. "Have you accepted the Xkit Guy as your Lord and Savior?" wrote one user. "Can we vote for the Xkit guy for president yet?" asks another. If the Tumblr Staff are a mythical demon that plagues the Tumblr fandom community, the Xkit Guy is the folk hero who does battle with them.
[5.3] The original, and most famous, Xkit Guy was Atesh Yurdakul, a college-aged software developer from Turkey who single-handedly maintained the browser extension from 2011 to 2014 (Romano 2013). More recently, the "New Xkit" browser extension has been maintained by a rolling group of volunteer developers based out of the New Xkit Developer Blog. Both Xkit projects have been open source, with repositories on GitHub where any Tumblr user can get involved ("FAQ," New Xkit Extension Blog 2020). Both iterations of the browser extension were released free of charge and are supported only by volunteer labor and user donations.
[5.4] Xkit is not the only browser extension in Tumblr's history. Before Xkit, in the early days of Tumblr, Jeremy Cutler created the "Missing e" browser extension, which was active around 2010 to 2011. Although Xkit had more general functionality, there have also been more specialized browser extensions. The still-active Tumblr Savior browser extension (developer: bjornsta0r), for one, was created specifically to allow users to blacklist certain tags. However, here, I focus primarily on Xkit because it is the most widely used and discussed browser extension and has provided the widest range of features and use-cases.
[5.5] Unlike The Staff, who are viewed by the Tumblr community as outsiders, the Xkit Guys are insiders to Tumblr's fannish communities, both the fandoms that inhabit the platform and the fandom of the platform itself. Both Yurdakul and the New Xkit developers maintained their own Tumblr blogs as well as the official Xkit blog, indicating that they had experienced the community as Tumblr users, not as outside developers. The New Xkit Developer Blog's header image features a meme referencing the web comic Homestuck (one of the largest fandoms on Tumblr) and the blog icon is a picture of Ryoko, the main character of the anime Kill la Kill, clearly signaling the developers' fannish affiliations. In contrast to The Staff's "Hello, fellow kids" style of talking down to Tumblr users, the Xkit developers' comfortable use of Tumblr's fannish vernacular (e.g., their competence with labels like "the Blue Hellsite" or "The Staff") marks their connection to the community. The original Xkit Extension often included jaunty little jokes and references to Tumblr culture—for instance, if you used the extension to download music files from Tumblr, you would get a pop-up with the text "You wouldn't download a car," a reference to a popular Tumblr meme mocking moral panics around piracy (Romano 2013). Unlike The Staff, whose new updates clumsily disrupted fan communities, the Xkit developers incorporated their view from within the community and their intuitive knowledge of how fans actually used the site into their design choices.
[5.6] Video game modding provides a useful comparison for understanding the Xkit Browser Extension. Video game modding is a phenomenon extensively studied by games scholars, in which players of a video game create their own code that runs on top of the code of the video game and modifies the display or mechanics of game play (Taylor 2006; Kücklich 2005; Kow and Nardi 2010; Postigo 2010; Banks and Humphreys 2008). Video game mods range from superficial (e.g., changing the skins of a video game or the display screen) to deep structural changes (e.g., creating new levels or characters). However, as T. L. Taylor (2006) has argued, even the most cosmetic mods can open up new modes of play—for instance, a mod that displayed a player's teammate's health bars on their own screen was essential for allowing large groups to participate in coordinated game play in World of Warcraft raids. Mods are usually created by volunteer developers from video game fan communities—though most game scholars do not explicitly use the word "fan"—and these developers' position as fans allows them to understand how other fans interact with the game and to provide features that those fans need in order to play the way they want to (Postigo 2010; Kow and Nardi 2010; Banks and Humphreys 2008). We can think of the Xkit Browser Extension as a mod, only for a social media site rather than a video game.
[5.7] Like video game mods, Xkit's modifications of Tumblr range from cosmetic to deeply functional. Some of Xkit's features make minor adjustments to the layout of the web page—such as moving the reblog button from the top to the bottom of posts. Other features run much deeper: the original Xkit allowed users to filter or block posts and to download audio and video content posted to Tumblr. Xkit can create new features or pave over unwanted official features: one of the most canonical uses of Xkit is to revert Tumblr to a more familiar version of the interface when the developers have pushed a disruptive user interface (UI) update.
[5.8] In addition to providing new features and allowing customization, Xkit is beloved for fixing glitches in the Tumblr interface. Especially between 2010 and 2016, the Tumblr Staff were notorious among fans for pushing shoddy updates that would break the site for days. This constant stream of glitches (in addition to the community's penchant for heated argumentation and absurd content) earned Tumblr the exasperated moniker "The Blue Hellsite." Tumblr users would frequently circulate screen caps of dashboard glitches with sarcastic comments like "Yahoo thought it was a good idea to spend $1.1 billion on this website" or "I love this functional application."
[5.9] The severity of Tumblr glitches is evidenced by the nearly encyclopedic memory of past disastrous updates that the site's community shares. In a representative interaction, one user wrote, "Does anyone remember in 2014 when Tumblr just randomly deleted blogs for 3 days straight?" to which another responded, "can't wait for the day the staff just randomly delete Tumblr itself [sic]." A search on Tumblr for the words "tumblr update" or "staff update" will usually yield pages of scrolling before a single positive post appears. In this context, Xkit's power to revert the UI, effectively undoing a broken update, was a godsend for users (figure 3). Like the video game mods described by Taylor (2006), which transformed unplayable games into playable games, Xkit is a mod that sometimes was necessary to make Tumblr literally usable.
Figure 3. Tumblr fans frequently make jokes implying that the Tumblr website is unusable, or unbearably garish, without the help of Xkit.
[5.10] Beyond basic usability, many fan communities rely on Xkit to provide affordances that make the site usable for fan activities. For instance, one of the main uses of Xkit is to filter tags. While Tumblr has an official feature that allows users to block particular preexisting posts, it does not contain a good, dynamic way to blacklist topics in order to block posts the user has not anticipated. Because Tumblr has no way to mark and hide spoiler text, open fannish discussion, for instance, of new episodes of a show risks spoiling fans who have not yet caught up. Xkit helps to avoid this situation: using Xkit a user could block "Doctor Who season 7," allowing them to browse for content on earlier seasons without risking encountering a spoiler. Major fandom blogs frequently post tutorials lauding the virtues of browser extensions and teaching fans how to use them to filter spoiler tags. Filtering tags is also a basic accessibility feature that allows users who suffer from post-traumatic stress disorder, eating disorders, or other mental disabilities to block potentially triggering content, allowing these users to participate in Tumblr's fan communities (Leetal 2019). When Tumblr failed to provide these tag blocking features after years of complaints from users, the developers of Xkit and other browser extensions like Tumblr Savior stepped up. These extensions became so widely adopted that its now common to see posts ending with the words "tagged [topic] for ts [Tumblr Savior]."
[5.11] The Tumblr Staff did not embrace browser extensions like Xkit and Missing e—but they had very little technological power to prevent users from making use of them. Browser extensions run on a user's own computer, so it was nearly impossible for the Tumblr developers to block them without creating extra work for themselves or accidentally blocking other browser extensions that users rely on such as screen readers. The best Tumblr could do in 2010, when they tried to block Missing e, was to add a popup to the site that politely warned users of the dangers of using unofficial software when it detected they were using a browser extension and to plead with users to turn the extension off (Valinsky 2015).
[5.12] Legal control, however, was somewhat different matter. The Missing e browser extension was eventually discontinued after Tumblr threatened legal action against the developer, Jeremy Cutler, if he did not remove a series of features that they argued were in violation of their copyright or put undue traffic on their servers (Warren 2011). Cutler, disheartened with the forced compromises that had gutted Missing e, decided to abandon the project; he wrote, "I am going to see what I can do to fix it up…Prune the features they don't want that I didn't feel right about including in the first place. If it's much more than that, I have my doubts about whether it is something I will continue to pursue" (2011).
[5.13] Though Tumblr has not directly threatened legal action against the more recent Xkit, the fate of Missing e has certainly influenced the current developers' understanding of their legal position with regards to Tumblr. In a Q&A about whether they were planning to add a follower tracker feature, which would alert users when another user had unfollowed them, the current Xkit team responded that "There is not, and will never be"; similar features had gotten Missing e and the original Xkit extension in legal trouble with Tumblr (Mod April 2018). As in the fan fiction communities described by Abigail De Kosnik (2016), and the modding communities discussed by game scholars (Postigo 2010; Kücklich 2005; Kow and Nardi 2010), Xkit exists under a tenuous, untested legal IP arrangement—Xkit's fannish producers keep themselves in line based on their own folk understanding of a presumed precedent as to what the Tumblr Staff will let them get away with.
6. The precarious Playbor of social media modders
[6.1] Both fan scholars and game scholars have debated and challenged the questions of labor, resistance, and exploitation as they apply to fan works or video game mods. On one hand, transformative fan works or mods can be empowering to fans, allowing them to insert themselves and their desires into the text, fix problems with it, and even oppose the official rights holder's power over it (Taylor 2006; Tushnet and Coppa 2017; Pande and Moitra 2017; Jenkins 2006). On the other hand, fan works can be a form of free labor, where the rights holders reap the economic benefits of fans' creative work without the fannish creators seeing a cent of it (De Kosnik 2009; Stanfill 2019; Stanfill and Condis 2014). These angles are not necessarily contradictory—Stanfill and others have explored how certain fan labor may be both empowering to fans and exploitable by corporations (Stanfill 2019; Stanfill and Condis 2014). These same debates and difficulties of categorization also apply to the Xkit Browser Extension.
[6.2] At a glance, we might think of the work that the Xkit Guy does as unpaid fannish labor for the Tumblr corporation. The work performed by Yurdakul or other Xkit volunteer developers does not actually look different from the work done by software developers employed by Tumblr. Like the video game modders described by Yong Ming Kow and Bonnie Nardi (2010), the Xkit Guys do not just create new features; they also do market research by listening to Tumblr users and collecting data on what features those users need. Xkit does not just save the Tumblr developers the time of developing new features themselves, but it also saves them even needing to bother to find out what users want. Furthermore, we might recall that these volunteer developers also are performing basic maintenance work to fix the site UI when it breaks—work that many fans think Tumblr should be hiring its own developers to do.
[6.3] The Xkit developers and other users frequently describe the work that goes into the Xkit extension as labor. In their explanations of why they decided to abandon their respective projects, both Cutler and Yurdakul explained that working on the browser extension stopped feeling like a leisure activity and started feeling like a job: "Hobbies and crafts you do in your spare time should be enjoyable. When they start feeling like work, they're tolling their own death knell," wrote Cutler (2011). Other Tumblr users also recognize the work of social media modders as labor—not to mention labor that Tumblr financially benefits from. One blogger, outraged at Tumblr taking legal action against the Missing e extension, argued, "The features they want to take away are the very ones that make the product great: Why are you not hiring this person, Tumblr, or at least taking some of his ideas and building upon them? He made your product better!" (Short Form Blog 2011). The Tumblr community is full of bloggers shaming the Tumblr Staff for not being as good at maintaining their own website as the Xkit Guy is. It is clear, then, that Xkit is seen as at least as high quality as any official feature of the website created by a paid developer, and yet the Xkit development team receives no money for it. Viewed through the lens of the "produser" (Kücklich 2005), Tumblr's benefiting from the Xkit extension is an audacious outsourcing of basic software development labor to unpaid volunteers.
[6.4] However, Xkit is also a tool that allows Tumblr users to seize control of Tumblr's design decisions, maintaining the affordances that are central to their communities, even in direct opposition to the Tumblr Staff's wishes. Using Xkit to revert to previous, less commercialized versions of the platform can be a powerful tool against Facebookification. Browser extensions allow user-developers to add features that they have repeatedly requested—features that meet the needs of users—even when the official staff fail to provide them. For instance, the ability to download audio and video files is an essential feature for fans to produce and circulate fan vids, fan cams, gif sets, and remixes (Jenkins 2013; Freund 2018), which Tumblr users have been requesting for years. Tumblr has historically refused to add the feature because of copyright concerns. However, the Xkit developers, who were not bound by fears of corporate liability, could provide this affordance that fan communities needed. Because of their position as Tumblr users, the Xkit developer teams were more in touch with fandom and had both greater ability to sidestep official systems and greater will to implement features that users requested.
[6.5] Working outside of Tumblr's payroll, the Xkit developers have even added features that go directly against Tumblr's profit motives, protecting fandoms from monetization. The New Xkit Extension, for instance, has an anticapitalism mod, which acts like a specialized ad blocker against sponsored content on Tumblr. Updates to this extension have especially focused on countering the "stupid f%$!@ing video ads" with auto-playing audio, which users find most disruptive (Mod Nightpool 2016). As is suggested by the name "anticapitalism," Xkit takes a more aggressive and oppositional stance than your average ad-blocking software. The developers of this extension are not just helping fellow users block ads as a convenience but are also taking a stance that they do not see Tumblr as a space that should be milked for profit. Thus, Xkit rejects a fundamental part of Tumblr's business model—selling users' clicks and eyeballs—in order to favor fans' visions, not the developers' visions, for the site.
[6.6] The developers of Xkit certainly see themselves as working in opposition to, not working for, the Tumblr Staff. The description of the new-xkit-discussion Tumblr in 2018 simply read "@staff fight me in the pit." Users may similarly view their use of Xkit as a radical act of resistance against the Tumblr Staff. The Xkit Guy has no illusions of wanting to, as Postigo (2010) would put it, "mod to the big leagues." The radical power of the Xkit Guys to make resistant code comes explicitly from the fact that they are not paid by Tumblr—they do not work for the owners of the platform but for the fannish communities that inhabit it. Seen in this way, Xkit is an oppositional tool that allows users to seize control of the site, make it accessible to groups who would otherwise be excluded by the affordances of the platform, and even punch back against attempts by Tumblr to profit from their data and eyeballs.
7. Code as transformative work, and as resistance
[7.1] In "Star Trek Rerun, Reread, Rewritten," Henry Jenkins (2006) describes fans writing fan fiction in order to "rescu[e] the show from its producers, who have manhandled its characters and then allowed it to die" (55). In this same spirit, Tumblr users created the Xkit extension in order to rescue their social media platform from The Staff. Like fans opposing Star Trek's producers, Tumblr users felt that The Staff, concerned only with money and not the health of Tumblr's fan communities, were mismanaging an IP that mattered to them. Like the fans of Star Trek, members of the Tumblr fandom claimed moral ownership of their object of fandom, taking management of the platform into their own hands by creating their own versions of it. And, like the Star Trek fans, Tumblr users took it upon themselves to fix problems with their object of fandom that the rights holders refused to.
[7.2] The Xkit Brower Extension, then, operates as a transformative fan work of the Tumblr platform. This claim does not feel so extraordinary when we note that video game scholars have acknowledged video game fans making fan works in the medium of code for years (Taylor 2006; Postigo 2010; Küklich 2005; Kow and Nardi 2010; Banks and Humphreys 2008). When fans of a book find problems with it, they can create fix-it fic in the medium of text (Goodman 2015; Jenkins 2006; Driscoll 2006; Russo 2017). When fans of a show want to get in on the storytelling, they can create fan vids (Freund 2018). And when fans of a piece of code, such as a video game or social media platform, have problems with that piece of code, one of the tools available to them is creating fix-it fic in the medium of code—such as a browser extension.
[7.3] In many ways, social media mods operate like other fan works. Social media mods are created by fans in order to provide things for their community that the original text will not. Just as a fan of a TV show who is upset that the showrunners killed off their favorite character might write a fan fic where that character is still alive (Jenkins 2006), the Xkit developers reacted to Tumblr killing off interface features they used by creating a fan work that revived them. Social media mods patch holes in the original text that the rights holders refuse to. In the same way that a queer user might find a text inaccessible to them due to its lack of queer characters and write their own (Russo 2017), Tumblr users who recognized that Tumblr did not have the necessary accessibility features for disabled users wrote their own. Social media mods help to legitimize versions of the source text approved by fans and delegitimize those that are not. In the same way that a fan of Dune might write fan fiction that ignores any canon established after Frank Herbert's death (Goodman 2015), Tumblr fans could use the Xkit Browser Extension to revert to older, more canonical versions of the Tumblr display.
[7.4] The case of Tumblr and the Xkit Guy demonstrates just how important the affordances of social media platforms are to fannish users, and the extreme lengths to which those users will go to maintain the affordances that their communities depend on. It also demonstrates the potential power of fan works to protect those affordances. In the digital space of the internet where, as Lawrence Lessig (1999) tells us, Code is Law, most platforms are not democracies but oligarchies, where the laws are laid down by a small group of developers who might be deeply removed from the communities that call these platforms their home. In the same way that fan fiction or fan vidding allows fans to speak their minds even when they have no power in the media industry, the Xkit Guy's "outlaw code" is a way for fans to claim a voice in platform owners' decisions and turn the code base into a two-way forum.
[7.5] I hope that my work on social media fandom and social media mods has demonstrated that many of the frameworks for thinking about fan works or modding developed in fan studies and game studies have broader applicability for understanding relationships between platform owners and platform users in general. Social media modding is a unique form of digital fannish resistance and participation, and a novel strategy that allows platform users to push back against what can feel like platforms owners' stifling and inescapable control of digital spaces.
8. Acknowledgments
[8.1] Thank you to Henry Jenkins, T. L. Taylor, and Louisa Stein for your guidance and invaluable feedback, and to Eleanor Graham and Linus Glenhaber for the editing help. And, of course, an enormous shout out to all Xkit Guys, past, present, and future—seriously, can we vote for you for president?