Tuesday, December 02, 2014

The unreal space of censorship - Protocols vs Companies


I'm kind of fascinated by the technical approaches to community "censorship" - in quotes because the word has taken on too many connotations to use casually.

Twitter has just announced new tools to discourage harassment, obviously in light of GamerGate. In the announce, they say "We are nowhere near being done making changes in this area." Along with Facebook's constant revision of privacy tools and policy, it's clear that controls over communication are being fought out in the realm of service provision (rather than, say, being user-controlled). The masses of users are, in all cases, totally dependent on the people providing the service.

I'm not saying that's a bad thing. It's just a fascinating thing given the history of this realm of control. It is fascinating having grown up with mailing lists and IRC as a collection of yardsticks, for instance.

In the "olden days" as my son would call them, you didn't have service providers. (Well, you had AOL. I'll come back to that.) You had protocols, and the protocols let anyone set up their own space. Sometimes there are spaces within spaces, such as IRC channels inside a particular IRC network. The protocols were vital - by being open standards, anyone can jump anywhere that uses them. Protocols encourage liquid connections. This is important.

Under normal protocol rules, the spaces that interactions happen in are real and not real; they can form and unform in a moment's notice. They are simultaneously everywhere and nowhere - easy to be established forever, and equally easy to be destroyed on a whim. Real ghosts.

Contrast that with Twitter. Twitter is a "thing" in itself. It is not a Twitter protocol. Twitter must survive because people need to make money from it. Twitter cannot be cloned nor destroyed. It must be preserved.


The cypherpunks mailing list was deliberately chaotic. Email is inherently non-authenticated, and the list added to this by attempting to preserve pseudonymity through a series of remailers that masked sending locations. Censorship - blocking - was purely the preserve of the receiver. If you didn't want to read missives from someone, it was your own filtering - usually a 'killfile' - that handled it.

It's possible to spend all day talking about killfiles, but the essential thing is that they were yours, no-one else's.

On IRC, things are different again. Anyone can set up a channel, or a chatroom. The "founder" of the channel has deity-like powers ("operator" access) over who can be or speak in the room, and can extend powers to others. From this simple set of rules, a system of reputation forms - those who gradually earn trust in the channel may eventually receive ops powers - or not, it depends on the way in which the channel was set up, and runs.

However, it is not the IRC protocol that sets those rules, or even the IRC network. It is the channel itself. If you don't like the rules, create your own goddamn channel.

I'm not saying any of these are good, or better. I'm just saying this is how things are in other places.


Twitter has a responsibility to its users that a protocol can never have - it is the provider, the police, and the politic. It defines how you interact as much as whether you can interact. It sets the terms. It determines the functions without a vote. How much does it define you?

Think back to AOL, which runs off the same lines, the same model as Twitter. It is the policed space that attracts a certain trust in its users. Without that trust, it is nothing more than a technical spec. Its business model is trust, and letting that slip results in debacles similar to Habbo Hotel in 2012. For a company, trust is everything.

But if that trust becomes the raison d'etre for a technical service, what does that mean for its users? Do they trust themselves any more? Are they disempowered? Or is the force wielded by the service they use something that they could never hope to impersonate anyway? What does it mean if they can't vote on the rules, or even work they way into the ranks, IRC-style?

You can blacklist but you can't whitelist. You can leave but that's it, you're out.

Either way, you're on your own.

No comments: