I'm working on a collection of networked games, around the central theme of a lockstep deterministic network model with 'delay frames' (between acting directly on user input and progressing actual shared network state).
- Neon Swarm, a take on the classic lemmings game, with multiplayer versus and coop modes
- Serpents, a snake game (where you eat food, grow a tail and need to avoid this growing tail), with inertial movement and multiplayer
- Spirits, an homage to N++, where you work together to get someone to the exit, while avoiding enemies, figuring out how to open doors, and so on
- Pilots, a multiplayer asteroids battle with homing missiles
Each game solves the network delay problem (the problem of providing immediate feedback to user input and hiding the fact that actual changes to shared state are delayed) differently, and it has been very interesting to work through a bunch of different approaches to this.
If anyone else here is working on multiplayer network games, I'm very interested in setting up a regular "play each other's games" session.
The idea is that regularly playing with other game developers will help develop a kind of 'scene' (where you get a group of people together who make work in public but really aimed at each other, pushing and unblocking one another to become bolder and better at an accelerating rate, as described here: https://www.henrikkarlsson.xyz/p/scene-creation-engines ).
It's still very much a work in progress, with just two games. One is a version of the classic game "Lemmings". The other is a very minimal asteroids style inertial ship duel.
The games both use deterministic execution in lockstep across network participants, and provide interactive gameplay and smooth execution even if situations with up to quarter or half second lag times, and the methods used for this are then perhaps one interesting aspect of the project.
In each game, the local player sees their own controlled entities at current positions, and other player entities at historical positions.
In the lemmings game, the complication is that lemmings depend on their own world changes (e.g. they need to stand on a bridge piece they just built). So there is a kind of fork and merge mechanism that enables local and remote world changes to be eventually consistent.
In the inertial duel game, homing missiles have a deploy phase in which they transition between different update time frames, with exactly the same sequence of updates applied on each machine, but with these updates accelerated or slowed down to achieve the time frame synchronisation.
There is an offline part with a tutorial to go through for the lemmings clone, which gives a flavour of the thing. The networking part uses a dedicated server and is currently invite only, but if anyone is interested in trying this then please shout and I can send out server credentials..
Recent discussions about artificial intelligence safety have focused heavily on ensuring AI systems remain under human control. While this goal seems laudable on its surface, we should carefully examine whether some proposed safety measures could paradoxically enable rather than prevent dangerous concentrations of power.
The text you get when you click share should indicate which puzzle you solved, in some way (e.g. like the number in "Wordle 247 3/6").
Having some kind of graphic in this sharing text would feel nicer, I think, as well, somehow.
(Could be just sequences of green blocks for each word, basically just showing how long the words you made are.)
The concept of diminishing returns, perhaps, together with the corollary, by which I mean the fact that you can often get easier profits or make progress a lot more easily by switching to some less explored avenue.
When considering selling points, everyone seems to be focusing on the advantages of being able to run traditional games without console hardware or delays for installation, but no-one seems to be talking about how something like this could fundamentally change the constraints on multiplayer game design. I wonder what the server side looks like, and whether anyone is working on MMO type products that are designed specifically for this setup..
I don't know exactly how that bug bounty program works, but it feels like there may be some fundamental issues with bug bounties in the context of open source projects.
For example, what is there to prevent someone from introducing a bug into an open source project, first of all, and then subsequently claiming a bounty for identifying and/fixing it?
(Maybe this can end up by forcing more careful pull request auditing for security critical projects, but that seems like a big ask..)
You can try this here: https://locksteparcade.com/Client
Games included in the arcade, currently:
- Neon Swarm, a take on the classic lemmings game, with multiplayer versus and coop modes
- Serpents, a snake game (where you eat food, grow a tail and need to avoid this growing tail), with inertial movement and multiplayer
- Spirits, an homage to N++, where you work together to get someone to the exit, while avoiding enemies, figuring out how to open doors, and so on
- Pilots, a multiplayer asteroids battle with homing missiles
Each game solves the network delay problem (the problem of providing immediate feedback to user input and hiding the fact that actual changes to shared state are delayed) differently, and it has been very interesting to work through a bunch of different approaches to this.
If anyone else here is working on multiplayer network games, I'm very interested in setting up a regular "play each other's games" session.
The idea is that regularly playing with other game developers will help develop a kind of 'scene' (where you get a group of people together who make work in public but really aimed at each other, pushing and unblocking one another to become bolder and better at an accelerating rate, as described here: https://www.henrikkarlsson.xyz/p/scene-creation-engines ).
If you are interested, let me know!