Editor’s Note: I’m sharing my research on how to build better online communities by studying internet history. Welcome to all 30k readers of last week’s post on Usenet!
Join the mailing list and follow me on Twitter for more notes and app updates.
Part 1: The Next Great Online Community
Part 2: Usenet — Let’s Return to Public Spaces
If you have a lot of users out there on the Internet and you don’t know who they are — how can you define the edges of who belongs to the community, and who doesn’t?
— Raph Levien
When you hear the word Advogato, you probably think of my two favorite things: avocados and cats (gatos in Spanish).
Me? I think of an online community from the late 90’s 🤣
It was a place for software developers to share their thoughts, known for being an early introduction to blogging and online communities.
It never went viral or amassed a huge userbase, but there’s a lot we can learn from it. For me, the biggest takeaways from Advogato were:
It was explicitly considered an experiment
It actively explored ideas around trust metrics and reputation systems
It had a fully transparent governance system
IMO, more community software builders should think of themselves as social scientists — observing what’s been done, and trying new things.
Rather than try to emulate others or succumb to the venture capital craze, Advogato’s creator (Raph Levien) followed his intellectual curiosity. He went on to try something completely unique for online communities of the time: build a trust-based community.
Not only that, but he constantly refined (and at one point completely changed) the way trust was measured within the system. He was truly a servant of the community’s interests.
More social software builders taking an experimental, trust-based approach is something I’d like to see.
Why?
Before continuing, let’s take a moment to talk about two terms that are often confused: reputation and trust.
Reputation is defined as: “the beliefs or opinions that are generally held about someone or something”. In other words, reputation systems measure what other people think of you based on how you’ve behaved in the past.
Trust, on the other hand, is the “firm belief in the reliability, truth, ability, or strength of someone or something.” Trust-based systems measure the expectations people have about your behavior in the future.
Trust-based systems are hard to find these days, while reputation systems are everywhere: Reddit karma, Amazon seller ratings, Yelp stars, etc.
In the world of online communities, the most infamous reputation system is Slashdot’s — with users that moderate comments, and users that moderate the moderators!
Unfortunately, trolls and spammers love sites with reputation systems because consequences tend to arrive after the crimes are carried out — they’re reactive.
Over and over we’ve seen the signal-to-noise ratios on reputation-based sites suffer as more people join. Like clockwork, they become a victim of their own success.
Editor’s Note: Slashdot also suffered from being sold, repeatedly, to companies with interests misaligned with those of the community.
One of the hardest parts of building a trust-based system is making it (to use the words of Advogato’s creator) “attack resistant”.
Basically, if someone with the most devious intentions comes knocking at your door — how easy would it be to carry out those intentions?
The first thing users will try is to create multiple fake accounts — so at the very least your system should be able to handle Sybil attacks.
Editor’s Note: Sybil Dorsett was famously one of the first reported cases of dissociative identity disorder — also known as multiple personality disorder.
Keep in mind: “attack resistance” doesn’t equate to being capable of catching rogue users 100% of the time. It means making it as difficult as possible to for someone to take actions that cause the community harm.
Advogato’s system did this — and more. It went beyond “block” and “mute” buttons to harness the most powerful influence on user behavior: trust.
It would’ve been interesting to see how it held up under the pressure of high user growth!
Adam Bosworth once said:
The currency of reputation and judgment is the answer to the tragedy of the commons
and I don’t know if he’s right — but there’s only one way to find out!
For more on how I plan to incorporate trust-based reputation into the app I’m building, and for more notes like this — subscribe and say hello on Twitter.
I’ll leave you with a treat from the Google Talks archives — a post-mortem of Advocado:
Spread love! ✌️
This is super interesting, I'd never heard of Advogato! The tiny social network I built explicitly leverages trust as well (among other bits and pieces, users can follow and/or trust others; only users you trust will see your private posts), so I'm quietly pleased I'm following in an old tradition without being aware of my actual predecessors. I suppose there aren't actually that many ways to skin a cat when it comes to trying to build microsocieties?