So I have an announcement. I sent an email to my work saying that I wanted to be able to focus on my work on taking advancing the federated social web to the next level. Read: better security / abuse resistance, richer interactions, virtual worlds.
How will I be funded to do this? I don't know. I do know that the state of the political world scares me enough that I feel I have to do this. If you want to support me: https://www.patreon.com/cwebber
Wish me luck. More updates soon.
@cwebber virtual worlds?
curious to see how that would work
I would think federation would be too slow for online games, so would it be different servers that just federate their information between each other
Let me clarify what "virtual worlds" means. Those who were around in the 90s may remember MUDs and MOOs and etc, which were massively multiplayer games before that term existed. In a sense, social networks today are degenerate forms of this design: we can talk to each other, but we aren't immersed in a sense of *place and interaction*.
The system I intend to build (Spritely) will allow you and your friends to program your social worlds together. Build it, dream it, make it.
If you want more hints as to what I'm thinking, I gave a presentation on this which contains some of my thoughts earlier in the year https://dustycloud.org/misc/mmose.pdf
Here's the crux though: I think making a decentralized social web that's hard to censor but focuses on protecting users is *critical* for the survival of freedom and autonomy of everyday people, *especially* in this political environment.
But what's the most common self-hosted server on earth? It ain't fediverse software, it's Minecraft. Why? I think the appeal of building a world with you and your friends is very appealing.
If we can make it very easy in-world to be able to program the entities you're interact with as you walk around and talk, that could be a big win.
The first engine that will be introduced will be for textual worlds with *some* amount of interspersed images, like oldschool 90s MUDs and etc, but with a bit more support for point-and-click interaction so new users aren't confused.
It could take weeks to model a cool robot and rig it to behave. It can take minutes to type up the equivalent text version.
@amydentata @andyc That said, once the text version works, I will probably look at building a top-down 2d engine world into the system also using the Liberated Pixel Cup assets and style guide http://lpc.opengameart.org/static/lpc-style-guide/index.html
Sekret: when I helped run the LPC years ago, this was part of the long-term con.
@clacke This is part of the theory behind Spritely: the dream to build a distributed massively multiplayer secure gaming environment might not succeed, but that's okay because it will almost certainly result in very interesting things. Flickr, Google Earth, much of object capability design, Python's Twisted framework... all of these things started out with the aim of building such a massively multiplayer game.
So I think even if the game bit doesn't succeed, we'll end up someplace interesting.
@craigmaloney Yes I have thoughts on how to do that.
"Spritely" is an overarching umbrella for a series of projects and demos I'm going to release, eventually culminating in Spritely itself, so some of the first demos won't have this, but as we get closer I hope to have this be a key component.
As you probably guessed, my work on the Goblins actor model library for Racket is aiming towards this.
Imagine if we ran fediverse software *as a game* that people played... that would give motivation for people not currently interested in the fediverse to join. Kids would play it, teenagers would play it... and now imagine if we hook it up to Tor .onion or I2P addresses, etc. This could help normalize the behavior of using those systems as much as https is normalized today. When everyone uses it, that can help protect politically at-risk groups, activists, etc as well.
@cwebber how would this system deal with griefers/trolls/abusers when everyone is behind anoymizers/tor etc?
also, see content fingerprinting, linguistic analysis, for identifying users based on usage patterns rather than specifically indentifying information
@sivy I have comments on that. I think that if the fediverse and actual anti-abuse works, it effectively has work on the same level as if most participants were "anonymous". Consider that even now anyone can spin up a node and start sending you messages. How to deal with that?
One approach is to not assume that everyone can equally message you by default and that you have "one inbox". Instead, there may be different inboxes that you hand out to different people depending on trust.