Comment by sneak | Hacker News Reader

Comment by sneak | Hacker News Reader

Parent Full thread

The idea of preemptive banning of federating with other instances is extremely common in Mastodon-land. This means that unilateral actions by admins, invisible to their userbase, restrict what their users can read, or who can read their users.

It would be fine if the users opted into this, but it happens silently and arbitrarily by admins, oftentimes based on speculation or gossip, not even real abuse. It’s all of the worst of tribalism, polarization, guilt by association, and preemptive censorship (regardless of whether or not admins are “in the right” by censoring messages between x and y flowing through their own machines).

It’s also not easy or reliable to migrate your account between instances without losing your followers, and none of the server implementations yet support virtual hosting, so you can’t migrate hosts while keeping your own domain/handle.

There are real problems there, and casually dismissing the major censorship issues in the ecosystem doesn’t begin to solve any of them.

You can't force user's eyeballs to read the bytes you ship to their computers unless you want to go full Clockwork Orange.

Some people want to exercise the rights over their computers (pick any ideology, FOSS included) and don't want certain bytes shipped to their computers. Who cares the reason.

Some people don't have the time, energy, money, and technical experience to exercise their rights of byte-shipping in a competent manner, so they carefully delegate that power to someone they trust. And some want to join in a community that is purposefully run this way. To categorically paint this use case as "insidious" ("silently", "arbitrarily") is in denial of these peoples' real needs.

Forcing peers to accept your bytes with the assumption that they must examine it with their own eyeballs in order to overcome a zealous interpretation of "censorship" is blatantly disregarding the humanity in a peer and their real needs.

It isn't contradictory or incompatible to say that people shouldn't be forced to do anything, and also simultaneously believe that censorship, especially the silent or invisible kind, is bad.

Would a web host performing MITM on an HTTP connection to alter or redact your blog posts be bad? After all, it's their hardware...

Instance banning is neither silent nor invisible. Every Mastodon instance has an About page (no login required) listing all instance bans and reasons, anytime. I would be in agreement with you about silent/invisible censorship, but that's not what's going on here.

This is a categorically different problem than MITM.

> Every Mastodon instance has an About page (no login required) listing all instance bans and reasons

There are instances which require an account in order to see the bans (cyber.space). There are instances which do not list bans at all. There are instances with made up reasons of banning made up instances (mastodon.art). Even that flagship instance lists incorrect reasons for removing instances (claims that certain instances shared illegal content when said instances do not allow any form of illegal content).

In addition most mastodon instances do not disclose their policies via AP. See for example https://fediverse.network/mastodon.art/federation

You're right, my mistake. In some cases it is not transparent.

However, this is not a systematic censorship problem, unlike centralized services with opaque policy language and a complete boot out the door. People are free to run their own instances or have multiple accounts across different instances.

Whether you think they're correct is irrelevant to the question at hand. Freedom of speech and association means you're free to not federate/talk to those problematic instances, and maybe you'd be much happier for it. On the other hand, not being OK with it and trying to fight for transparency means you're trying to externally force these communities to be run in the way you want, which may be received well, but not always b/c forcing unwanted change is exactly the opposite point of Federation: communities will be built the way their members want to build it. Like the real world, some value transparency and some don't.

It's one thing to argue specific bans about specific instances and disagree on the other party's interpretation; it's a totally different claim to say that the entire system is corrupt with opaque censorship.

Mastodon != Fediverse

Do not put words in my mouth. I only replied to the point regarding transparency in your post. I really don't care about the rest.

I'm sorry! Based on context, I understood your post to refute mine to support sneak. And sneak and I have had heated debates about the Fediverse before, and you've stumbled into the latest one. :)

In the future, it would definitely help me and others understand your motivation better if you could even include one more sentence in your communication like "Just here for a correction: some instances are transparent..."

I will strive to be more charitable.

>claims that certain instances shared illegal content when said instances do not allow any form of illegal content

That can be a simple issue of jurisdiction. Mastodon.social is hosted in Germany (IIRC), so they have to adhere to German law. That means, for example, while hatespeech isn't strictly illegal in the US, it certainly is in Germany, it even has a fairly good legal definition. Or take the Japanese instances, which aren't well federated or have media-bans because of differences in media legality. And lastly it can also be simply the case that the instance is not moderating (ie, they write 'no illegal content' but do not care).

Both the statement that an instance shared illegal content and that the same instance was banned for illegal content can be true at the same time.

Why does anyone have a requirement to talk to your instance? What if, due to your ideological stance on issues, I don't trust that your server won't decide that mine is distasteful in some way, and that you'd cause your server to disrupt mine in some way by flooding it with messages, for instance? I have legitimately had this happen and it nearly brought down my instance. Some script-kiddie decided to fuck around and spam my account on an instance I run with about 2000 followers in the course of a few seconds. This almost killed my server entirely. Why should I have been forced to interact with their server on the grounds of free speech?

Why does it matter if any instance decides they don't want to associate with you? It doesn't affect your ability to use the service beyond not being able to interact with folks who probably don't want to talk to you anyway.

Forcing someone to make their server software talk to yours is just as much of a "Free Speech" infringement, if not moreso.

> Why does it matter if any instance decides they don't want to associate with you? It doesn't affect your ability to use the service beyond not being able to interact with folks who probably don't want to talk to you anyway.

It prevents people on that instance who explicitly want to follow me from doing so.

It also prevents me from following people on that server from my primary account on my homeserver, even if those people explicitly want the whole world to be able to read their public messages.

Both of those are undesirable interference between mutually-desired communication by Alice and Bob, by Mallory.

Then they, and you, can find another instance where you can mutually talk with each other. You are, in fact, allowed to have multiple fediverse accounts for different purposes/groups of people.

This doesn't address my point at all that you cannot argue that my server is somehow obligated to process bytes from your server.

Tools that disobey their end users are unfit for purpose. They're bad tools, and should be replaced with useful ones.

Your peers' routers are allowed to drop your packets, but nobody is arguing that that's good or beneficial. There is a difference between "within rights" and "good".

While I don't agree with you, there is an interesting discussion to be had here about digital "commons" (something like the "public utility" concept in the US), I think.

If a company with a near-monopolistic network effect (Google, FB, etc.) censors speech or who-can-talk-to-who/see-what on their platform, it seems that most folks agree that this is bad, whether or not they're willing to do anything about it.

So, instead, we have decentralized services (and semi-decentralized ones, like Mastodon). At what point does a Mastodon community operator's decision to censor speech or who-can-talk-to-who/see-what on their platform become problematic? When a community achieves a large size? If new community members aren't made aware of the censorship? Is the difference between these communities and an ultra-ubiquitous one like "having a Google account" or "being connected to friends on Facebook" a difference of degree, or a qualitative one?

It's simple to argue an extremist position of "all speech between any set of parties must not be suppressed for any reason, even by the parties themselves", but I don't think most people want to live in that world. Similarly, it's pretty hard to isolate a line past which an operator of a community-service should be held to a different standard of conduct because of how ubiquitous/depended-on their community is, but a lot of people seem to think that this line exists.

Reply on news.ycombinator.com