Skip to content

Instantly share code, notes, and snippets.

What would you like to do?

Tent is not a Tent developers' chat room

(December 2012)

A Tent user, “Alex”, recently posted:

It appears that tent is becoming filled almost entirely with posts about tent. Think I'll wait until a user-friendly app or two is released and then have another crack. Not giving up on it, but at the moment it holds no appeal except for geeks.

And it's true. Tent, the social networking protocol I wrote about in October, has been growing its ecosystem and attracting more (and more active) developers, which is great. I've been thrilled to see all the #tentdev posts.

But Alex isn't the only non-developer bemoaning all the technical and meta-posts. In a sense, we're already suffering from our own success. More people talking about Tent should be a good thing, and the fact that it is instead an annoyance and a problem, suggests we need better tools to filter out the noise.

Hiding replies

Twitter, by default, hides replies from your feed unless you follow both the author and the recipient. I'm not aware of a Tent client that does the same. TentStatus set the precedent of showing you everything those you follow post, and client apps like Zelten, Tentia and Tentmonstar have copied this behavior.

It's been argued that this is a feature, not a bug. You get into random conversations and find new people, say some, precisely because you're seeing Alice's reply to Bob, even though you don't follow Bob. Tent gets more people talking to each other instead of just broadcasting.

Which is true, but if you want to go in that direction, IRC does an even better job getting people talking to each other. The Tent always-show-replies behavior makes the global Tent network less like Twitter, and closer to a chat room.

Chat rooms don't scale, and they can't serve disparate groups of people. In the jubilation immediately after Tent's launch, we were all talking about Tent and so we were all interested in what everyone else had to say. It was exciting. Everyone was curious to listen in on every conversation.

That isn't the case anymore. Hiding replies by default would make Tent more manageable and help all of us see less of what we don't care about.

Developers: use a separate account for test posts

When I started experimenting with different Tent clients, one of the first things I did was to create a separate, free account specifically for testing. I made it very clear in the profile that it was a test account and that it belonged to me (linking to my regular account).

Posting “test” or “foo” or “post” is pure noise, and it wastes the time of everyone who follows you. It also contributes heavily to the impression that Tent is a developer playground, not a real social network. If you're a developer, and people follow you, they probably want to hear about new releases, progress updates, and opportunities for becoming beta testers themselves. They don't want to see “test post” every 15 minutes.

Future: Public aspects

Metcalfe's law states that networks become exponentially more valuable as they grow. But the idea originally referred to telephone networks, a technology for one-to-one communication. Growth can make social networks — which allow many-to-many communication — less valuable to their users.

Consider the reaction of twentysomethings when “old people” like their parents started signing up for Facebook. It was widely seen as a weird or awkward development, and it made posting on Facebook less fun. We may share anecdotes about drunkenness and sex with our peers, but we don't want mom and dad in on that conversation.

Nowadays Facebook addresses this concern with Groups, arguably a copy of Circles in Google+, which in turn may have taken inspiration from Diaspora's Aspects. You can limit your post to a certain group, and no one else will see it.

As developers, we have a more subtle problem, though. When we post about open-source code or our Haskell epiphanies, we want everyone to be able to see it, even if they don't have a Tent/Facebook/whatever account. But we don't want to spam our non-developer friends with it.

Kevin Kleinman calls his solution public aspects, crediting the original idea to Minecraft creator Markus Persson. Essentially you have a group that is public and anyone can join without needing your approval. You add your mom to your Family group. Open-source hackers add themselves to your Code group. You yak about protocol changes to the Code group, and everyone is happy.

It's a little too early to think about public aspects seriously since, well, Tent doesn't even have regular private groups working yet. But we should keep the idea in mind. I already have this problem on Twitter, and Persson, who doesn't want to spam his friends and family with Minecraft updates, was complaining about it on Google+ more than a year ago. People have multiple identities and interact with different groups of contacts differently; social networks shouldn't force us to jam it all into one place. The lack of public aspects on existing social networking sites presents an opportunity for Tent to do better.

Why I don't use anymore

(February 2013)

I used to think the protocol was a cool idea and a great project. See, for example, the unchecked enthusiasm in my previous post on the subject. Not anymore.'s developers have a weird preoccupation with preserving all your content forever. For example, if you edit your Tent profile, people can still get all the old versions and see info that you removed. And last time I checked, not only was this the default behavior, there was no way to opt out of it, either. And they claim this is a feature but nobody I have talked to wants it.

In emails, they hyped the (not fully developed) privacy settings in their software, and said you should use those. So... if you make an error in judgement, they just throw you to the wolves? For all the criticism Facebook gets, if you remove info from your FB profile, it's legitimately gone. Not so on, and this is a huge trap everyone should be wary of.

I also followed the Tent software's development for several months after launch. It broke frequently and progress on making it do what it was supposed to was slow. Plenty of features were promised ("we'll implement X next week") but never delivered even months later. Ever since the initial release, the ratio of shipping code, to talking about how cool it'll be, has been extremely low.

Overall, I don't think the developers really understand the problems social software solves. The undeletable version history just goes to illustrate that misunderstanding. They also seem unable to draw a line in the sand and say: "Our project does this, but it will not do that." They seem to think anything and everything should be a app — no focus.

I don't mean to discourage anybody from building apps as a learning exercise, or a proof of concept — it could be great for that. But if you are looking for decentralized social software that solves real problems and has a shot at relevance, then you, and I, are going to have to keep looking.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment