As people have probably guessed, a lot of the demos you've seeing out of me (the magent paper, Goblins, and yes, even Racktris) are Spritely-related. For those who don't know, Spritely is my longer-term project for distributed online social environments... imagine a fediverse MUD or etc, but with much stronger security and privacy properties than the Fediverse has.
I'm going to keep doing this for a bit, churning out demos; they're going to get more visibly close to the real system as I go.
I feel embarrassed about the slow pace; it would be great if I could work on this full time; development would be much faster. I'm in a reasonably good place in that I have a nice $DAYJOB where I do work on some related problems and with great people (though I do hate working with javascript).
I've considered asking for money for my Patreon or something to see if I could replace that income, but I dunno. Maybe I need more to show before such a thing is even worth exploring.
@cwebber Put the Patreon page up, goddammit. You're one of the people responsible for the future of the Fediverse and I'll gladly throw money at you because of that.
@phoe Oh, I have a Patreon page, I'm just not advertising it really: https://www.patreon.com/cwebber
I think it would take hitting $1500 / month for me to be able to totally work full time on this stuff though, and that seems like a lot to raise. In the unlikely event that somehow I got that much coming in per month, I could switch to working on building cool fediverse-converging-with-distributed-games things full time.
That seems like a lot to raise though.
@phoe I might be able to do it if every time I release one of these demos I accompanied it with a "like this? donate to my patreon!" type thing or something, which is I guess a thing people do
And since I'm working I'm of course thinking about the demos I will be writing next... probably one of the following:
- A minimal ActivityPub server tutorial?
- Add support for actors distributed across processes and machines to Spritely
- A minimal chat client and protocol in Racket (demo goal: practice for for the MUD'ish AP client)
- Finish that damn petnames paper already
- An actual petnames "contact list"
- Maybe one of the above explicitly in conjunction with I2P or Tor Onion Services
Oh, and:
- Figure out how to get live hacking ability into Racket
- Port of the MUD code I wrote in Guile to Racket (you can see more on the video on http://www.gnu.org/software/8sync/ )
- Add an object capability secure language to Racket for secure remote code execution (so players can contribute NPCs and etc to the server safely)
🇳🇱
@freemo BTW if you want to see how that distributed system could possibly work, this paper involves a lot of ocap jargon but it has most of the right ideas https://ieeexplore.ieee.org/document/4480750/?reload=true (can't find the non-paywalled version anymore though)
@cwebber Nice, going to check that out too.
@cwebber Awesome, thanks so much. I'm glad I wont have to design all this from scratch. I'm all about standards. Can I help in your efforts in any way.
@freemo Not yet but keep watching here. I hope to get up some demos soon that may request that people experiment with the user-generated-content stuff. That'll be a good time to jump in.
@freemo Reading that paper, reading the ActivityPub spec, and reading up on object capability security are good steps in preparation for now though.
@cwebber Yup, ActivityPub I've studied already.
Going to attempt a deep dive on those papers and ill hit ya up with my feedback.
@freemo Can you read lisp'y code btw? If so, you may find this paper on object capabilities from a language perspective helpful http://mumble.net/~jar/pubs/secureos/secureos.html
There are many ways to do object capabilities, and the above paper explores a programming language approach
@cwebber I can read lisp/clojure yea. Though I rarely develop in it. I should be able to grok that link, just added it to the reading list for today.
@freemo Cool. So take those ideas and apply them to the actor model and you get where I'm going. (E is an example of a language that does this, and ActivityPub is an actor model system too, but not explicitly an ocap one)
@cwebber I think after i read through this stuff ill be able to get into your head space. I'd like to have a convo on that maybe tomorrow if your around. Just gotta get up to speed first.
@cwebber Great, I'll wait with baited breath and help where i can. Since my game is text based it can likely adopt new ideas more quickly than a graphical game. So great place to experiment I'd think.
@cwebber Ohhh this is so cool, I was actually working on federating my own MUD game, we should collaborate or at least share notes (reading your doc now).
My MUD code is here (screenshot attached of in game map I recently improved).
@freemo Mudsync is free software, you can see a demo video of it at http://www.gnu.org/software/8sync/ and the code is at https://notabug.org/cwebber/mudsync/
But the plan is next to move to a distributed, federated "MUD" like game https://dustycloud.org/misc/mmose.pdf