Mourning Mastodon
Realizing the death which preceded the Universe
What follows is an attempt to come to terms with a loss that I have found myself grappling with over the past several weeks; the focus of this loss being the fledgling software Mastodon, a user of which I have been for quite some time. For the uninitiated, Mastodon is a free and open-source server software designed for communicating with and participating in the large, nebulous social-media network colloquially known as “the fediverse”; in very recent times, it has found itself the subject of a surprising level of journalism, hailed frequently as the (enlightened, doomed, sometimes both) successor to Twitter and its ilk. This characterization is not wholly accurate: Mastodon is not, strictly speaking, Twitter’s successor so much as its replacement: Inheriting none of Twitter’s business models or strategies, it provides a non-commercial alternative to Twitter’s service without replicating its model. In this it is not alone: Mastodon is but one in a number of implementations of the open network protocol known as OStatus, and Mastodon servers are able to more-or-less seamlessly communicate with other servers in the network.
In part as consequence of the aforementioned media attention, and in part because of a few high-profile users who have spread knowledge of the network through word-of-mouth, the OStatus network and Mastodon in particular have recently experienced remarkable growth — at the time of writing, instances using the Mastodon software numbered over 1000, each hosting anywhere from a single user to tens of thousands. If the OStatus network is a “fediverse,” Mastodon servers themselves certainly make up a large galaxy, each instance its own planet with any number of users orbiting around.
On beholding this incredible expansion, one might expect to experience a sense of wonderment and mystery — the excitement and joy of an unknown frontier, or a newfound popularity. That I should instead find myself stricken with melancholia suggests that this transition has been less than benign. Almost overnight, Mastodon has found itself shattered, fragmented, beside and unknown to itself, and in this turmoil, I fear that certain elements have disappeared from view; indeed, I find myself questioning whether any semblance of them still remains.
Of what do I speak? Put simply: the very spirit of the project itself. With no explicit goals or mission statement, the Mastodon project was instead strongly conditioned by the circumstances and community which it cultivated and served, an arrangement which, due to its growing popularity and changing demographics, exists no longer. What fate this will spell for the software moving forward is unclear, but that this original community was largely queer and deeply motivated by ideas of social justice only makes its loss that much harder to bear.
In proclaiming the death of Mastodon’s soul, however, I knowingly enter into a genre of criticism which is frequently characterized by bitter conservatism, in which any and all growth and change is cast aside as a foul taint, tarnishing the divine inspiration and ideals of its original conception; it is not my intent to follow this path. The changes which have come to the Mastodon platform in the past few weeks have been both good and necessary, and I am of the opinion that further changes will be required moving forward. But acknowledging the strength of these changes does not require admitting that they came without harm. To the contrary: I am of the belief that Mastodon’s present iteration came precisely at the cost of the community which birthed it, that this expense was both egregious and unnecessary, and that the extraction of this toll is doomed to continue unless definite steps are taken to rectify this situation in the future.
I will present this argument in three parts. In the first section, I will analyse the history of the Mastodon project in an attempt to better understand the forces from which it emerged and the community on which it relied. Far from being a complete piece of software which sprung whole and clean from its creator’s mind, the history of Mastodon is one of negotiation, turmoil, progress, and change. By better understanding the contingent nature of the software’s features, we will be able to more accurately characterize the driving forces behind the project as a whole.
In the second section, I will turn a more critical gaze on this history, deconstructing the power imbalances inherent in its structures and making note of their impact on the project’s development. Many of these inequities have gone heretofore unacknowledged, and have been made clear only in retrospect after the circumstances surrounding them have changed. Nevertheless, their impact remains.
In the final section, I will lay out a series of demands regarding the practices of the Mastodon project, moving forward. Having identified the importance of the project’s original queer community and acknowledged the oppressive structures under which they have and continue to operate, these demands aim at achieving justice for both this community in particular, and other marginalized groups as well.
I will stake these claims not on the basis of being a developer — for, although I have in small ways contributed to Mastodon development, the sum total of these contributions are rather minor and hardly worthy of note — but as a user, who has been on the network since nearly the beginning, which is to say, November. Consequently, my critiques will not be technical (which is not to say that there are not technical critiques to be made), but rather will reflect the social, political, and cultural contexts from which Mastodon emerged and in which the software now resides. It is my hope that the analysis contained herein will help software projects like Mastodon continue to grow and thrive, with an eye to the ethical ramifications of their procedures and to the confines of their success.
Despite Mastodon being only half a year old, I find myself far from the first author anticipating its demise — with networks like Ello and Peach still fresh on everyone’s mind, those quick to point out its flaws are dime-a-dozen, usually taking the form of journalists who have little used, or taken the time to understand the history of, the software. These being above-all-else tech journalists, their criticisms are generally technological, bemoaning Mastodon-as-software’s lack of features, inability to scale, or the informal, non-corporate nature of its development as the Achilles heel foreshadowing its inevitable downfall.
This is not the argument I am trying to make, not the least of all because it isn’t true. The technological side of Mastodon is, by all available metrics, thriving and full of life. In the past week alone, the Mastodon repository on GitHub has merged 258 out of 298 active pull requests, and closed as many issues as were opened. Over 100 users pushed nearly 300 commits, and the lead developer’s Patreon currently brings in a monthly wage of over $3,000. The most important marker signalling the success of the project, the dedication and commitment of its userbase, has held strong since the program’s inception, and it seems clear that, for the foreseeable future at least, Mastodon-as-software is here to stay.
But we are not tech journalists (or at least, not all of us are), and need not confine ourselves to so restrictive a definition. Indeed, there is no reason for us to assume that the success of a project is reducible to that of the technologies it produces, or even that one might not thrive while the other falters. But what, then, comprises a project?
Whatever else it may be, Mastodon-the-project is the author from which Mastodon-the-technology emerged, and so perhaps by considering the nature of the latter we might be able to shed some light on the former. We can immediately note two things: First, that as a technological product, the Mastodon software is necessarily a product of labor, and thus, by extension, of the conditions of labor which lead to its development; second, that as a social network software, Mastodon necessarily anticipates (and, through its deployment, creates) a community, as without an (at the very least, imagined) community, the project of building a social network becomes unintelligible. As a free and open-source project, these two definitions are deeply intertwined with one another: The labor from which Mastodon was born has and will always draw from and be conditioned by the community which it serves. And while I am not attempting to make the claim that the Mastodon project as a whole is reducible to simply a particular community, laboring under a particular set of conditions, certainly these things form an essential part without which the project would be no more.
The usefulness of this definition becomes apparent when we compare Mastodon to other implementations of the OStatus protocol — for convenience, GNU Social. Like Mastodon, GNU Social is a part of the fediverse, and from a technological standpoint, the two pieces of software are not so different — by necessity, they are similar enough to be more-or-less interoperable. However, the cultural context and conditions of labor of these projects differ so wildly that the two seem to perpetually find themselves in conflict with one another, with GNU Social users frequently complaining about Mastodon’s Patreon monetization and “trigger-happy” approach to censorship, and Mastodon users lamenting GNU Social’s light stance towards moderation and “we were here first” demeanor. Neither project would happily be conflated with the other.
Consequently, when I suggest that Mastodon has died, or is soon to perish, I am not predicting that the technology or software is going to fail, or even that its development is about to cease. Rather, I am pointing to the fact that the community and conditions of labor under which the software was created have now been supplanted, perhaps irrevocably, and in a manner so complete and extreme that very little of the original project remains. That Mastodon as we know it today descended from the fledgling social network of the past is indisputable fact; that they are one and the same is hardly a given. Rather, it seems to me as if the modern Mastodon has killed and replaced its former self.
But let us not speak so readily in abstractions. Proceeding again from the hope that by working backwards from Mastodon-the-software we might learn something about Mastodon-the-project, I will now enumerate those aspects of Mastodon which seem to me to define the service, and trace these back in an attempt to understand the conditions which brought them to surface. Surely, if there is a thing which we can call “the Mastodon project,” it will be found there.
I will categorize these into two groups. The first contains those features which have been a part of Mastodon since more-or-less “the beginning”; we might think of these as Mastodon’s conditions of birth:
- Support for federation/OStatus
- The (global/federated) public timeline
- General account and post actions (posting, replying, following, liking, reblogging, etc.)
- The prime instance (mastodon.social)’s strong stance against Nazism and harassment
- Account blocking (at the time, a “soft block”)
- 500-character posts
- Support for sensitive media
The remaining features were instead added after the project was well underway, and might be considered its life developments:
- Per-post privacy settings
- Content warnings
- Locked accounts
- Private posts
- “Hard” account blocking and account muting
- The local timeline
- The welcome modal
This second category is the one with which we are concerned. While the initial conditions of the Mastodon project can make for interesting study, it is the way that the project has grown and evolved from this point that best characterize its qualities. It is the nature of software development that the historical contexts from which features emerge are often erased, and for users who joined Mastodon during or after its latest boom, the idea that there was once a time in which post privacy was determined on a per-account, rather than a per-post, basis might seem difficult to fathom. So allow me a moment to elucidate on these developments here.
When I first arrived on Mastodon on November 23, it in essence a better-moderated, less private version of Twitter. Accounts could be made either “public” — which meant that their posts would appear on the public timeline and in hashtag searches — or “private” — which meant that they wouldn’t. At this time, replies were not excluded from the public timeline results, and massive reply-chains formed as users jumped freely into ongoing conversations happening on public accounts.
This system, surprisingly, worked for the first few days, until November 25 rolled around — as some might remember, the day Fidel Castro died. Political flamewars engulfed the site. The very next day, the project’s lead developer pushed a commit to hide replies from the public timeline, and shortly after made post “privacy” a per-post setting. Now that it was easier to do so, and to prevent further flare-ups in the future, the community largely agreed to the practice of keeping politics off the public timeline.
Of course, this agreement didn’t last. As political developments occurred around the world, some users (especially newer ones who had not been around during the Castro debacle) would inevitably feel the news too important to keep behind closed doors, and after a number of similar but smaller disruptions the community worked itself out a compromise: Politics, suggestive/lewd content, and other potentially unsavory material would be allowed on the public timeline only insofar as it was presented in a manner such that the casual observer would not be disturbed; the ROT13 cipher quickly emerged as the preferred mechanism for this, and bookmarklets and conversion tools abounded. Within a month the public timeline was filled to the brim with garbled, ROT13'd text, to the extent that getting started guides of the time had to include a section to explain them. When content warnings were added near the end of January, it was not only as a more robust solution to the problem but to prevent new users inevitably being greeted with grkg jevggra va guvf znaare.
Meanwhile, problems with harassment meant that locked accounts and private toots, features initially resisted on account of their difficulty federating, could no longer be avoided, and implementations for each were developed near the end of December. Mastodon’s “soft block” became hard and a “mute” functionality was added (much later) as a less-extreme alternative. After a massive schism around race and politics led to the creation of awoo.space and drove many POC off the site, “report” functionality was added and the local timeline was finally implemented to allow finer community control and practices moving forward. Prior to the massive influx of new users in April, users were welcomed to the site by volunteers and through hashtags like #welcome and #introductions; only after the rate of new users became too large to handle was a welcome modal introduced to assist with onboarding.
From these stories one might notice two things: First, that far from being the product of enlightened and planned foresight, new features were added to Mastodon generally after their moment of need first arose, and even then frequently only after a good deal of strife; second, that in every case these features were the result of deliberate requests from members of the community, many of whom directly contributed to the push for solutions by writing code and submitting pull requests through the project’s GitHub site. It is no exaggeration to say that the community which Mastodon served for its first six months made it what it is today.
What of this community? What was it like? Well: “Welcome to mastodon.social, here’s a copy of the Communist Manifesto and your fursuit,” or so the (oft-repeated) saying went, underscoring the deeply leftist, kink-positive and often furry nature of its core base. Often these users were lesbian, gay, bi, or pan; frequently they were trans and/or nonbinary; many dealt with disabilities; many were victims of harassment. Certainly, the most vocal and most frequent of Mastodon’s unpaid contributors seemed to invariably fall into one or more of these categories.
To say that these users formed the flesh and blood of the project is no overstatement: They structured its community, they funded its development, they volunteered their own labor. Insofar as Mastodon is presently successful, it is so through their (largely unacknowledged) continual support and efforts, at once critiquing and building up the program to reach new heights.
To say that Mastodon has died is to say that this body has ceased to be maintained.
The power inequities inherent in the Mastodon project were not initially made clear. Mastodon’s community was largely queer and, consequently, so was its funding, and so were most of its software contributors. Queer ideas surfaced, queer code was writ, queer technology resulted. Many users looked into the software and saw themselves reflected back. There did not seem to be any cause for alarm.
Of course, in retrospect, it becomes clear that only some queer ideas were implemented, only some queer code made it into production, and the image which resulted reflected not the community’s desires as a whole, but rather the choice selections thereof, those pickings which had been deemed suitable for mass-production. At the time it was easy to miss: This creator’s code got merged! This community member’s idea finally got made! Every week brought new victories. For a project that had only been in existence for a few months, the solving of the remaining issues seemed to come down to only a matter of time.
Looking back, however, it is easy to see that not every idea was destined to be picked up on. Pinned posts, in addition to or instead of a larger bio space, were immediately identified as an important concern, both for users who needed to provide in-depth information about themselves or their needs, and for artists who wanted to include a work or link that could then be quickly viewed and reblogged. Support for displaying pronouns and other metadata in their own section on user’s account pages was proposed, and a mockup created, in late November. A robust system for federated private posts, ideally in addition to unlisted posts which are nonrebloggable, became a request the moment private posts were implemented, as the current situation locked users onto the same instances as their friends. Ever since the first 500-line post appeared in the public timeline, users have asked for a way of collapsing overlong posts; a pull request to address this problem was proposed but not merged. Adding custom alt-text for images, in a similar manner to Twitter’s image descriptions, was identified as an important feature for improving website accessibility. Supporting quote-reblogs, or at least providing notifications when a link to someone’s post was posted, was considered by many users an important safety feature, while self-untagging was determined an important usability concern. Support for lists and mutuals-only posts were proposed as enhancements on Mastodon’s existing privacy settings, and instance-level blocks were seen as a requirement for controlling who had access to a users’ data. Support for multiple accounts, account migration and data exporting features were proposed as ways of preventing users from being locked into a single instance. Every one of these features remains unimplemented at time of writing (and the list goes on).
Naturally, not every idea is necessarily a good one, and the fact that not all suggestions have received implementations is not necessarily evidence of nefarious forces at work. Mastodon is a largely-volunteer open-source project with limited time and resources, and it is understandable that certain ideas would take time to reach fruition. However, I bring up such long-standing issues because they begin to hint at a disconnect between Mastodon’s original base and those calling its shots — between its body and its head, so to speak. Although the queer community on Mastodon made up a significant portion of its early adopters and have contributed to the project in meaningful ways, they have never held any real decision-making power. The fact that it often took near-catastrophe for their suggestions to be heeded is a testament to this fact.
The consequence of this structure was a situation not unlike Twitter, or Facebook, or other larger platforms, in which users dealing with oppression — including those who are queer, but also including people of color, disabled users, and victims of harassment — were constantly engaged, consciously or not, in a battle for recognition and validation, both of themselves (who were frequently left out of the narrative when it came to Mastodon development) and of their needs and ideas. It seemed as though every new Mastodon feature needed a subsequent patch because it had been designed with too low of color contrast for users who are visually impaired, and this is but one example out of many.
Given their large numbers among Mastodon’s early adopters, this arrangement was workable for queer users of the site for some time — as they formed its largest and most influential community and both funded and contributed to the labor behind the project, they held some amount of sway over the product which resulted. (Not so for users marginalized along other axes, who were on more than one occasion driven from the site.) It was not possible to excise, or even ignore the queer community from the Mastodon project, because the queer community was the Mastodon project — the inspiration and unpaid labor of the project, anyway.
This is no longer the case. The recent influx of users to the platform has brought with it new contributors and an expanded revenue stream that has rendered the original nearly obsolete. Queer users could leave en masse without harming the project’s survivability, which means that the reciprocity of their relationship has been terminated — queer users still depend on the project, but the project no longer depends on its queer users. This is, undoubtedly, a dangerous situation.
It was also a preventable one. The relationship between Mastodon and its queer community has always been a negative one: They have worked together not for their mutual benefit, but to prevent their shared destruction. We can see this pattern in the list of ideas which received adoption and in the history behind their development and the network’s spread. However, there is no strong reason why this needed to be the case. Had Mastodon given queer contributors the ability to make executive decisions regarding the project, the community could have reached a place where it was no longer in peril. Had it implemented support for data exporting and account migration, it could have eased their dependence on the platform.
Instead, we have a diaspora of users, locked into accounts and instances and networks, a disparate but entrenched community on a platform which no longer needs them. For users who have spent months building up friendships and communities and networks of support, they simply have nowhere else to go. What was once the heart and soul of the project has now become a body ripe for exploitation, and unless a change in course is made this extraction is guaranteed to only increase in the future.
That the original queer body of Mastodon would eventually be overwhelmed may seem to some as inevitable, and indeed, it is an unresolved question as to whether maintaining such an exclusive base was even desirable, given the history of conflict and harassment on the site. But one does not need to chop down a tree to plant a forest. We can celebrate the growth and diversification of the Mastodon platform while maintaining an eye to the community which produced it, and ensuring that those involved receive justice.
What do I mean by justice? Simply this: that the community which produced and shaped the Mastodon phenomenon be recognized for their efforts, allowed to enjoy the fruits of their labor, and given the opportunity to continue to play a meaningful role in its development; furthermore, that they be granted these things even as, given the realities of demographics, they may never again serve in the majority. If justice cannot be served in this manner, then I ask that they be provided with the tools (data exporting and account migration features, at the very least) to liberate themselves from the platform and develop projects of their own, without losing in the process the labor and communities that they have worked to produce over these past months.
These requests are for the most part things which might rightly be expected for any marginalized population, and are linked to ideas of social justice in general. But the history of the queer community as the progenitor of the modern Mastodon project places them at a unique position from which to make these demands. If even they are unable to attain justice, the future for other communities appears bleak indeed.
As it stands, Mastodon is not well-equipped for serving disparate communities’ needs. There is but one version of the software, with no established means of extension; there are no advocacy positions or research projects by which communities can make themselves known; the development pipeline seems hardly attentive to or even aware of the specific communities which its product serves. These are all things which will have to change for the project to be able to accommodate anything more than its loudest majority.
And while I am not necessarily optimistic about the response that these demands will receive, we must make them all the same.