The self-hosted pipe dream

My ultimate aim, for my online presence, would be to be completely self-reliant. The aim is, objectivele, a bit of a pipe dream, since I'm well aware of the gigantic efforts that it would entail to actually reach a point of total independence, (and for starters, even for something as essential as email I am pretty sure that I will never make it) but I like to get my gains wherever I can.

The most obvious example (you're reading it now) is the choice to abandon external blogging platforms in favour of this self-hosted wok. For other protocols (such as IRC), “self-hosting” doesn't really make sense, but controlling your client allows you to control your experience and most importantly your backups (I have all the logs of all my IRC conversations, without having to ask them to anybody else). For some services this is not possible, although sometimes there are “bridges” that allow some degree of client control: for example, I use Bitlbee to connect to some of my instant-messaging accounts through an IRC-like interface, although the reliability of these “bridges” is severely limited by an explicit intent from their operators (e.g. Google) to limit interoperability.

In some cases, I have simply scaled down my presence, sometimes helped in that by the demise or downfall of the corresponding platform: this is for example the case for the now-long-gone FriendFeed (my idea of what could have been “social networking done right”, even if still on a proprietary, centralized platform), or Tumblr, which I only sporadically visit (most of the content being luckily accessible from other platforms or “followable” through other means).

Up until recently, the only remaining “significant” online point-of-presence for me has been the microblogging platform Twitter. I must say that even my on-platform presence has been sporadic for a long time, but had recently gained some weight, and each of my post there has been made with a “second thought” about the loss of control over my content. Much of it may be recoverable through the Twitter data export feature, but it's still a non-trivial process, and the fundamental implication about the loss of control remains even when workarounds are found.

(By the way, even for Twitter it's possible to set up a Bitlbee bridge, although given the extensive use of graphical elements the experience is far from being as smooth as with other services.)

The social network escape

I do not have the time or inclination to discuss the dangers of the centralization of social networks (especially when others have written more and better than me on the topic, and I may even link some of the relevant content from e.g. the EFF or Cory Doctorow here, after I find the time to collect it), but the work to create open and distributed alternatives has been going on for over a decade now, driven largely by the interest of individuals and groups worried about the implications of proprietary control of online spaces.

The most famous example at the time was probably Diaspora*, born as an alternative to FaceBook in 2010, that even reached a certain prominence in the news during a bout of «delete Facebook», but was hardly the first or the most successful (for example, the microblogging platform, alternative to Twitter, had already been active for a couple of years).

The coordination of efforts from separate groups, each dedicated to a specific aspect of the “modern”, “social” web (blogging and microblogging, aggregation, music and video streaming and sharing, discussion fora, etc) has led over the years to the creation of what is now known as the Fediverse, a “universe” built on the “federation” of individual entities. The development of common protocols (most notably the now recommended ActivityPub) and the growing maturity of the developed software has finally reached the point of (at least technical) feasibility for an alternative to centralized social networks, altough the question remains about the possibility for them to become a viable, and widely adopted, alternative to the centralized platforms.

Enter Mastodon

The recent bid by Elon Musk to purchase Twitter and make it private (“to restore freedom of speech”, but I will discuss the idiocy of the claim and of those who actually believe it in a different time and place) has brought back into the news the alternatives to it, and in particular the currently most popular Mastodon.

As with all Fediverse components, Mastodon is not a hosted platform in se (in the sense of a centralized website to which users register), but a software stack that provides a platform. Each installation of the software is an instance, and there are therefore multiple websites to which one can register to have a Mastodon account, similarly to how people can get an email address from different providers (their ISP, Google, HotMail, etc). And just like with email, Mastodon users can communicate with each other, follow each other's updates, etc through the common protocol, regardless of the instance they are registered with.

Not being a centralized service puts a barrier to entry on Mastodon, especially for people used now to decades of centralization: making a Mastodon account requires an active and conscious choice about where (which instance) to create it, with the associated burden of understanding the difference between instances, the fact that they each have their own terms and conditions (and possible additional restrictions on who may or may not register with them), and so on and so forth. There are “general” instances, both global and language-specific, that may be considered the go-to fallbacks, and are probably a safer bet (in term of reliability and permanance) compared to smaller instances: these provide, for the technically uninclined, the closest thing to an “optimal” situation outside of fully centralized solutions, similar to the larger email providers nowadays used by most people (the classic HotMail, Google's gmail, etc), with the benefits (and downsides) of the decentralized, federated model.

I personally don't think that the decentralized model poses a particular obstacle to adoption (despite the slightly higher barrier to entry), and I will discuss elsewhere the details on what may make (or fail) the future of the Fediverse, aside obviously from the FUD propaganda fueled by centralized services that are threatened by this model (hint: it involves the participation in the Fediverse of some high-profile accounts, possibly on their own instance in an official form: think for example of institutional accounts from the US or EU being on their respective or instances, in contrast to e.g. the unofficial mirrors from the proprietary platform that can be found through the instance).

(Edit: I just found out that the EU actually has an official Mastodon instance. That's actually pretty good news.)

In fact, from my “self-host-all-the-things” perspective, a much larger problem with Mastodon is that it's non-trivial to set up a personal instance: while it is possible, Mastodon is a bit infamous for being a massive resource hog with complex setup necessary even in the “reduced” use case of a self-hosted personal instance, to the point that it's frequent to find recommendations to try alternative microblogging platforms that still integrate in the Fediverse, most typically Pleroma.

As a result, I won't be able to consider myself commpletely self-reliant on the “social network” side of things yet, even while moving away from the centralized platforms.

Testing Mastodon

So yeah, I've taken the opportunity to set up a Mastodon account on a general instance. I'm not particularly worried about the future of Twitter with Musk at the helm (in fact, I doubt anything would change, and it even looks like the deal, that was assumed done, might actually fall through), but like in other circumstances, I've grabbed at the chance to stop putting off exploring alternatives to Twitter when the circumstances presented themselves.

I don't consider my current Mastodon account to be “definitive”, even though it will likely last for several years, as I don't see myself switching to a self-hosted instance anytime soon, although that would be the ultimate goal. In the mean time, I've made an effort to set up as much as possible in a way that would allow me to interact with the platform “on my own terms”: this includes setting up a Bitlbee bridge to be able to follow and interact with my timeline from an IRC-like interface (I must say that I'm not particularly impressed by its stability yet), and the adoption of a practical command-line utility that I've used extensively to search and follow various accounts from a variety of instances. Periodic, automatic backups of my stream are something that I intend to explore soon.

The process of migration from the proprietary platform to Mastodon will take some time, a transition period with permanence on both, and will probably result in a long tail (most likely reduced to the few really interesting “big name” accounts that won't switch or clone their presence across networks). Luckily, a lot of effort has gone already in the community to help in this regard: I've discovered a few services that make it easier to bridge Mastodon and Twitter, including a “wrapper” for Twitter accounts that are presented as if members of a Mastodon instance, and a service that should make it easier to manage accounts across different platforms (I haven't tested it yet, but it should come in useful while transitioning from the proprietary to the open platform). There are also ways to improve one's visibility across instances, such as this Italian bot designed specifically for this purpose.

I have not even made my first Mastodon “top level” toot (post) yet (sharing this article will probably be the first one), but I've already had the opportunity to interact with some users (unsurprisingly, mostly about the nature of Mastodon itself, as the influx of new people exposes doubts and perplexities about its accessibility, long-term viability, and the potential of the social platform on its own merits rather than just to host people running away from the proprietary platform for whatever reason or exploring the fad of the moment). The experience has been rather smooth, although I've noticed some minor issues already (most notably, the fact that you cannot set the toot language when posting from the web interface, the less aggressive/expansive behavior for embedded links esp. other toots —which may or may not be an issue, depending on the use case— and the fact that a fixed column size is used in the “advanced” web interface).

It will be interesting to see how the currently much lower traffic on Mastodon develops in the following days and months, if the migration pressure keeps up or dwindles (by general drop in interest or because the Musk acquisition falls through, if it does). It will also be interesting to see how well the platform holds up as the influx of new users puts strain on the software and hardware instances holding the network together (already the general instances maintanied by the Mastodon developers themselves have shown signs of “cracking under pressure”).

And who knows, this all might lead to better, more lightweight software and possibly more interest in making self-hosting more approachable.