@EsolangingFruit neither of those are fully fixed; a now-defunct nomic known as "einos" didn't have a fixed player list (rather, it allowed anyone to take game actions at any time); and in most long-lived nomics, it's rare for most of the players to be trying to win at any given moment
@NathanMerrill not once the rule change ability has been removed; however, a game with no potential to remove the rule change ability is typically considered to be not a nomic either
@ais523 That sounds like actions wouldn't be reviewed by humans. If so, did einos require humans to review actions or did it have some way to formally specify them?
OK....I could see this being possible: If you split the controller into two parts: The ruleset and the fundamental controller. The fundamental controller calls start() over and over(). If an error occurs, all players lose. If start() returns a player, they win.
then the ruleset is just a bunch of code that can be modified and recompiled according to its ruleset
@EsolangingFruit one of the most fundamental ideas in nomic philosophy is the idea of a platonic and pragmatic gamestate; the platonic gamestate is objectively specified and manipulated automatically as people perform actions; the pragmatic gamestate is what the players believe that the platonic gamestate is
most successful nomics need a) some way to cause the two to converge, and b) some way to inform participants of what the gamestate is
the longest-running nomics I'm aware of are Agora and BlogNomic (both of which I've been a player of in the past)
in Agora, there are players known as officers who publish reports on their belief as to what the gamestate is every week
if the reports aren't challenged for a week the platonic gamestate changes to match the report
this means that it's impossible for the platonic gamestate to get massively out of sync to what players believe it is
in BlogNomic they have a public document specifying what the gamestate is believed to be and game actions are made by editing the document, meaning that it's usually up to date
but it's not as secure as Agora's method because if someone edits the document illegally and nobody notices, it causes a gamestate divergence that can take a while to figure out and fix
however, BlogNomic also has frequent gamestate resets which help to avoid problems like that persisting too long
well, it's a sort of virtuous cycle; the reports are supposed to describe the gamestate, if they fail to and somebody notices it'll be corrected, if they fail to and nobody notices we go with whatever the believed gamestate is
voting on rule changes is relevant because the standard for a rule passing by proposal (more than half the players voting on it, for a basic rule) is easier than the standard for a proposal passing via intentionally incorrect report (none of the players objecting)
majority is easier than unanimity
and yes, the system is described within the rules and even changes from time to time
getting the gamestate changed "by ratification" (i.e. posting a deliberately incorrect report and telling people not to challenge it) used to be a very rare and contentious step reserved only for emergencies
it's tried more often nowadays but still doesn't succeed very much
@EsolangingFruit in its full form? Not for a while. In a simplified form? Definitely. What you need is 1. Game state and 2. A way for participants to add/remove rules. You'd probably need to write a simple language for that though
@Dennis I'm opening a mathematica TIO link I was using a couple hours ago, the font for the code is messed up, I can't see the code. I can copy, edit, and highlight the code but that's all?
@quartata depends on how easy it is to run, and how long the challenges are. e.g. for formic functions the controller needs no setup so everyone uses it, plus it's been running for a long time so everyone would have set it up by now anyway
a sophisticated physics engine, the ability for me to make future KOTHs on more sophisticated maps, the ability for anyone to watch or for humans to fight them
CMP: would a Hello world be valid if the language could not print the exact string? or would be using the language as a decision model language be preferred
