On Mastodon flaws, and decentralized services

Mastodon is the latest “Twitter replacement”. Yesterday I said on Twitter that something described as its “bigger flaw” seems no big deal to me. Here I explain why.

On Mastodon flaws, and decentralized services /img/mastodon-bigger-flaw-no-small.png

I had never heard of Mastodon until 2 days ago, when I wrote that yes, even Nazis can use it. Yesterday I found this comment on BoingBoing to a more complete article by Sean Bonner. The Boingboing comment ends with this quote:

“if I want to be @seanbonner@mastodon.social or @seanbonner@mastodon.cloud or any of the other instances then I have to create separate accounts on each of those, and there is no way to sync them. This also means that some other Sean Bonner can go sign up as @seanbonner@anothermastodon.instance…"

“This is the biggest flaw in my opinion because without the ability to claim your identity across an entire service there is huge potential for confusion and no way to embrace it as a home”

As soon as I read that quote (not the article), I instinctively reacted tweeting:

  • 1/2 @seanbonner & @frauenfelder, #mastodon “huge potential confusion” is the SAME that AFAICT never was big deal with #email, why bother?
  • 2/2 besides, that is a general issue with ALL open, federation-ready services. A fair price to pay for more freedom, less control, IMO

Bonner answered:

  • All the other successful federations like… email? And… that’s it. Nothing else has worked at scale. Did you even read my article?
  • email is private social media is public. Please try and pay attention.

After reading the whole article, I basically confirm my initial impression, but let me explain it better.

First, THE big, big problem with Mastodon, or any other federated service as they are today, is that they don’t automatically federate, that is connect, and stay connected, with each other. Bonner is absolutely right to say that “this is a huge problem”. Fixing it is a prerequisite for what I argue at the end of the post.

But this problem is how things are gone so far, not the only way they can be. Besides email (but for the same reason, in a sense), another successful federation that is working at scale is the World Wide Web. All websites can link to all others, and each web page has a unique address, thanks to DNS and URls standards. Yes, we do need some guaranteed interconnection among independent instances of federated services, but nothing more.

Second, I disagree that “Email is inherently private and social media is inherently public”. I know plenty of people who use email for their public activities much more than “social media”, which they reserve for chatting with friends and relatives. It may be more exact to say that email is inherently one-to-one, and social media inherently many-to-many. But that’s not the real issue, of course.

Third, the central issue, that is identity. Bonner writes that:

“Your email address is not your public identity where as your social media accounts often are. And while it’s true that no one would try to lock in the same name@email.server for every email host out there, it’s also true that there haven’t been massive lawsuits and fights over email addresses the way there have been for social media account usernames."

On one hand, the main, if not the only reason why there are certain lawsuits and fights is exactly because those are services with the wrong architecture: fully centralized, and controlled by one, not so accountable entity.

On the other, even on centralized services, the problem exists and has no solution. Only the FIRST Sean Bonner to subscribe to Twitter had the privilege to be @seanbonner there, for no other reason that happening to be the first. What about all the other, equally authentic Sean Bonners who had exactly the same right to get that account name in Twitter?

Ditto for me, of course. When Gmail started, by pure chance I was invited soon enough to secure marco.fioretti@gmail.com (which is NOT my main email address, by the way. This is). Here in Italy, there is a lawyer with my same first and family name. Several times a year, I receive totally unwanted, confidential email from his clients, who merrily type his full name.surname into Gmail. But that only happens because, in all these years, that guy has not set up yet his own website, business cards, letterhead… with all the correct contact info. NOT because the Web, email or DNS have some “fatal flaw”.

This is why I think Bonner is worrying about, if not perpetuating, a non-issue.

Real federation, yes. But I don’t really need an account with my name on every instance of any service like Mastodon, no more than I really need a “marcofioretti” subdomain on any possible website. Even if I could demand it, which I can’t. What for?

With the Web and email, every individual or company of the world has managed their identity in the same way. We walked into an open, decentralized service, enjoyed its openness, and dealt with its consequences without making a fuss: if you can, you register your preferred domain name. If not, the closest one. In both cases, you must add a short description of who you are. It’s only thanks to that that people discover without ambiguities if one is THE Sean or Marco they want, or not.

But even on services without the “fatal flaw” of Mastodon you must do the same! How do you know which of the many Seans or Marcos on Facebook, Twitter, etc… is THE one you’re after, without looking at their picture or “About” tab?

Paragraph added on 2017/04/08: if you aren’t convinced yet that certain worries about “unique names” are overrated, check out this story. There is no way unique names alone would have prevented those problems. The only solution would have been to always have other information TIED to the names of all those persons: pictures, profession and other details. Just accept it.

Fourth, the federated future. When walled gardens came, we accepted their way to communicate and declare our identity. But that is not good. If Twitter goes bankrupt, or shuts you down, your precious, so well “established identity” disappears. With federated services, it cannot happen.

Besides being a non-issue, the “confusion” Bonner worries about is the only way forward, if we want to get real control on our identities, and data. It has to be done right, of course.

In my opinion, the only real solution consists of:

  • just stop to worry about identities inside services,
  • get used to, and demand, instead, complete personal clouds (*), and working, automatic federation among them.

By “complete personal cloud” I mean a personal online server, with its own unique domain name, bundling its own instances of software for (micro)blogging, email, file sharing, instant messaging etc… which of course should be able to “automatically federate”, like email servers can find each other, with other software of the same kind on other personal clouds.

With such a server, everybody (both individuals and businesses) would have their own identity, once and for all (the domain name, e.g. “myname.com”) and automatically have accounts/addresses like, for example, email@myname.com for email, tweet@myname.com for microblogging, and so on.

Bonner writes:

“I’ve had people call me @seanbonner to my face… but no one has ever referred to me as my email address”

I want a world where that sentence becomes:

“I’ve had people call me seanbonner.com to my face… but no one has ever referred to me as my email address, OR Twitter handle, OR Mastodon handle.."

There is lots of work already ongoing on federating services as I suggest, from Solid to SocialWG, just to name two. That’s the way to go. Let’s worry about bundling and federating open protocols and services, not about “names” as such. Not linked to services, at least.

(*) my very own idea of the perfect personal cloud is here. You’re welcome to implement it yourself, if you like it and have the means, or to support any of the other proposals of the same kind.