Obscuritus.ca A nerd making a nerdy blog

Federated vs Global Communications

I've been working on this idea a lot, and trying to pull it all together in my mind has only really made it more and more and more deep, so I am hoping that perfect is the enemy of done, and if I write some of it, I can get at least part of it down.

I've been looking at Federated Friend Locking, Twitter, Facebook, and Quitter. These things are all important to how I am seeing social networking, which I am not sure is doing what we expect. As I've been reading, talking about thinking about social networking (I'd cite my sources if it wasn't just a mishmash of concepts and ideas, mostly from Singpolyma, Kousu (currently not on the web) and Randall Howard). I am currently thinking about how most of social networking is about screaming. The cult of the ordinary celebrity. Which, really, is part of the promise of democratic capitalism - that anyone can become powerful - which is awesome.

Part of the problem with social networking that people identify is that people scream about themselves a lot. Of course, that's part of society. But, at the same time, it's partly that we've moved a lot of our private interactions into the public sphere. The blurring between bragging about events, speaking about politics, protests, and family conversations. We basically are going through our lives in public - it's really hard to separate our identities. And we've built a lot of tools to colate our identities, it's hard to split our identities into multiple places - and it's hard to do anything in the private sphere. We're all beginning to have the issues that people always warned us that paparazzi caused for celebrity. Maybe we're a paparazzi culture, rather than a celebrity culture.

At the same time - my father and mother follow me on twitter. My RPG friends can see my github, and so on. Things default to public, you have a single identity, a single location, and a single stream for your thoughts. You can't believe something wrong, you also can't muse silently. Note the defensive tone of this - it's because it needs to be real - but it also needs to deal with my own weaknesses publically.

To make it worse - everyone discusses employers looking at social media. Unless a largish number of people self censor (which appears not to be the case) this is going to be useless. This is definitely just a scare tactic (I am perfectly willing people are willing to scare tactic themselves). This is basically just a statment that we have a system of unintentional viewers, and that unintended viewers are undesirable for our private lives (don't post pictures of what everyone agrees is fun, or else people wont hire you, since you once had fun).

This basically comes down to two main axes - control and discoverability. I think part of the problem is that most people are focused on control. The more we focus on control - the less we deal with discoverability. I, personally, think that federate services fix both of these problems perfectly, and I'd like to make the case for using federate services. Therefore, I am going to define both of these terms (control and discoverability) and how they relate to several services. Then - I am going to define federation and attempt to discuss the pros and cons of federation within the context of control and discoverability. Once I am done - hopefully this will be self evident.


When we are dealing with control in the context of social media - we need to discuss the facts of who owns data, and who gets to control interface, abuse, and advertising. For the most part, the administrator gets to be the person who specifies the interface (especially if it's a web interface), own the data (unless it's email), define what is abuse, and what abuse will be dealt with, and with who gets to advertise. For the most part, a lot of services define abuse as "advertising without permission" (since it cuts into revenue), nudity and possible profanity (since we're notably squeamish), and then anything that's harassing (unless it fits into a cultural narrative the admin supports, or is unwilling to take stances on). One may assume that the admin also gets to ban users, or ban external hosts.

The question of who gets to specifically control their own data, and the services they use is super important. This is very commonly discussed in the media when they discuss privacy. By giving your control to Google, that means that Google gets control of your data and therefore the ability to read your data. This is the core of why the FBI wanted Apple to decrypt a phone - everyone was discussing how this was the FBI asking the owner of a house for a copy of the key. And people kept stretching that metaphor in order to explain why it was bad. But really - the FBI was asking the BUILDER of a house for a key - and wondering why the builder didn't have a copy of the key. Since, the FBI expected Apple to be the controller of the data, rather than the owner.

Part of what is currently happening in the world of anti-harassment and safe spaces is when marginalized groups demand that those to whom they've surrendered power, give them safe spaces with which to protest, or accept power. Most notably this exists with nursing mothers getting angry that Facebook censors pictures of themselves nursing, or when people being harassed get annoyed that the people running Twitter don't understand their needs (note, these are emphatically needs, and if we are to surrender power to a group, we MUST demand they use the power we surrender to them appropriately).

Most of the time, people want to treat all online platforms as public, instead of in the control of very specific people and interests. They want the protections that come from either their own control, and the idea of a public space. However - as long as they do not control the spaces that they want to use for protest - they cannot properly have a voice with which to protest. This is another topic, but it seems to me as if many people are angry that they can't use the platforms of the powerful in order to fight the powerful. I really wish we could though.

Part of the reason why someone might sacrifice control is in order to allow ease of use. Not everyone is a sysadmin, and as someone who has successfully run a mailserver in the past, I am glad that most people don't choose to run their own mailserver. It makes sense that we want Google or Yahoo to run our mail - but we should also demand some level of control as part of that contract. Services providers not giving up some control, should cause us to move. Why it doesn't baffles me.


The next thing is that of discoverability. Depending on the tool we are using - we might want our content to be discoverable. Web indexes (such as Google or DuckDuckGo) really increased discoverability on the web - which can cause some of the "lost in the noise" sense of the web to go away. Big Data is focused entirely on matching ideas to people, and also matching ideas to people across identites. Part of my thoughts have been around the idea of discoverability of our communications. Discoverability is when your posts can be discovered by anyone.

A lot of people with problems with harassment (basically any marginalized group) have the problem that highly discoverable content is quickly attacked by harassers. Letting people get at their content quickly and easily allows people to attack them quickly - often for private jokes and workshopping ideas which aren't yet done. It's not exactly a space where ideas can be built and new dialogue can grow.

By giving people spaces where they can speak privately, you let them deal with concepts that aren't ready for the public stage. Now, part of the call for Safe Spaces is the call to have spaces where people aren't judge for their first thoughts. People need to have the opprotunity to say things they aren't proud of, and say things that are not public. The more private forums are discoverable, the more they are simply public forums being used to say private things. When people are critical of Twitter for making all tweets discoverable - they are misidentifying a public forum as a place they can express private opinions.

A lot of discoverable speech has the issue of where it is being placed. Speech on Facebook is discoverable by its nature - that's one of the goals of Facebook. I think that a lot of people want their speech to be discoverable since they think that what they have to say is valuable and that people are interested in them (I am not judging those attritudes as bad, but simply identifying the fact that most people do indeed like to be heard). However, this breaks down in a couple of ways.

The most obvious is that of the stalker/harasser, who can follow anything discoverable. Without a robust method of blocking someone from reading your speech, you wont ever be truely free of this problem. Of course, this means making most speech not default to public. For example, posts only going to members of your Circles in Google+, or protected Twitter accounts. If most twitter accounts were protected (and public accounts the unusual form), or most posts went to a default Circle, then people would be able to better control who saw their posts, and their block lists would be more meaningful. Of course, defaulting to whitelisting, does lose a lot of the value of these services (since discoverable communication is nice for starting more international conversations).

The other drawback to discoverability is the problem where people can always find someone who agrees with them. Breitbart News is a great example of someone who simply created a website, and started writing articles, and after a while had gathered a large enough following to be considered a news source. It's not that they have credibility, it's that they pander to groups, and the groups they pander to use them as an example of a news source that agrees with them. If everything is discoverable, then you can always find someone who agrees. Unfortunately, the problem with this is the problem with web indexes in general. As long as anyone can quickly and easily find any opinion, they can find a community based on that opinion. This is a deeply complicated problem, and much more suited to sociology than technology, and therefore isn't easily solved via federation.

Generally - discoverability is a matter of public vs private as long as we allow that there will be web indexes around that are very good at searching the web.


Finally - we get to the point of federation. Basically, federation is small services that connect to one another. Examples of federation include email, GNU Social, XMPP, and UseNet. Things like Facebook and Twitter are definitely not federated. Things like OwnCloud are also not federated. Basically - I define a federated service as one that is designed to be run in many places, conain many users, but interoperate seemlessly with other instances of the same kind of server. You likely don't run your own email server, and there are likely more users than just you on your email server, and yet you can send email to nearly anyone with an email address with the same ease as you would send email to people using the same provider. In the same way, XMPP is equally easy to use if you want to chat with someone on the same server, or on another server.

The advantage of this setup in terms of control is that the user gets to be as in control of their service as they want, placing the systems administrator as far from themselves as practical (many people use GMail, which goes what they need, at the cost of having Google have some control. Many people use their ISP's mail, which they pay their ISP for. Some will also host their own mailserver, meaning they pay in terms of time to run updates and configuration). Depending on the user, they will make the choice of where they want to put their server, and therefore, as part of that, how much control they sacrifice. Social networks like Diaspora or UseNet allow you to do this as well, where a user can host their own node, which can then get shared with larger nodes.

One of the advantages of this is that abusive nodes can be flagged and removed from the network (email even has RBLs to federate the flagging of nodes) allowing people to remove problematic people from the network, as well as allow per-user and per-server killfiles in order to slow abusive situations that could explode elsewhere. One of the important things about most federated services is that most people don't pull in the entire world at a time, meaning that it would take effort to discover messages from someone you weren't looking for. As much as we'd like to broadcast many of our messages publically - requiring someone to actually know something about what they are looking for, means that there is a higher level of work required in order to find targets for harassment, and a little bit of digging is required to find communities around your ideas. In fact, a lot of the time, in order to find members of a community you aren't already a part of you will likely need to interact with larger groups of people.

Basically - I believe that by spreading information out and making smaller communities, people will be able to make their own choices about how they will engage with a community, and also, need to engage with a larger amount of the commnity in order to truely be a part of it. By decreasing discoverability, federated services require more social churn, and more ideas to come up before the affirmations occur.

By bringing out social interactions into the public sphere instead of the private, we limit the ways that we can challenge our own ideas, expose ourselves to harassers, and give up the ability to curate our friends.