This is why I appreciate still being close high school LAN party friends close online--I can start a group chat and get a game going between each other, and maybe they can invite a friend of theirs to get the number of players up for a game. If you can manage to maintain these relationships then online gaming can be very rewarding without having to worry about the trainwreck you described.

This doesn't work with all types of games obviously, and perhaps all online games from here on out will have kernel-level protection. In the meantime, I think we will keep running UT2004 in compatibility mode (or in my case, Wine).

Did they ever release a version of 2k4 for linux?

I can play the UT2003 native version for linux just fine (doesn't seem to play nice with amdgpu, but the intel driver seems to be fine).

UT2004 has a Linux port, I believe developed by Icculus[0] but it's pretty long in the tooth. Requires libc5, an old version of SDL, and OSS for sound.

Regardless, it'll still run on modern systems if you acquire old libc binaries[1] (and preload them), use SDL 1.2 (possibly a patched version[2] which enables Alt+Tab) and start the game using aoss[3] for sound.

[0] https://icculus.org

[1] https://www.unix-ag.uni-kl.de/~deusser/UT2004/

[2] https://github.com/infertux/SDL-1.2.7

[3] https://alsa.opensrc.org/Aoss

I don’t think they ever did. I run mine in Steam using Proton without any configuration minus a widescreen fix, which has actually been a lot easier than trying to get it to work in Windows 10 (I use amdgpu)

It came with the Linux installer on the DVD, same as UT2003.

The fact is online multiplayer games are fundamentally broken. It depends on trust and that's rare because almost no one knows each other. Everyone's a stranger. It's just random people playing together. People find ways to cheat even on trusted platforms like game consoles.

People should be playing online with others they personally know and trust.

Fundamentally broken in theory. In practice:

- The percentage of people who want to cheat is not that high.

- Any cheat that spreads widely enough can be obtained by the developers and detected. Cheat developers can and do sell exclusive cheats to smaller groups of people, but fewer people using the cheat also means less disruption.

- With tactics like delayed ban waves, game developers can make cheating risky enough to create an effective deterrence, even if they don't actually catch all cheats.

- If all else fails, game developers can have players manually review other players' replays for cheating, like with Overwatch for CS:GO (not to be confused with Overwatch the game). So obvious cheating will be caught, and if people make their cheating non-obvious, well, that also makes it less likely to annoy other players.

These measures can't stop all cheating, but they don't have to; they just have to deter it enough that it doesn't unduly hamper most players' experiences. In practice, it seems like most games are able to accomplish this.

But these games commonly employ intrusive anticheating software, do they not? Regardless of its effectiveness, that counts as a large part of the genre being fundamentally broken.

Csgo isn't intrusive at all

That's pretty much true for real-time games. There are turn-based games that can be played on a company-provided server, and some extreme cases where games are actually decentralised (like blockchain chess). The open source RTS 0ad used to have every client compute every game state, but I believe that's no longer the case. Without smart contracts you fundamentally can still have decentralised computation of game state, and even preserve some asymmetrical knowledge by having players encrypt orders and share keys after all players have submitted orders. A decentralised game of Diplomacy could be implemented this way. Have each player submit a seed and use the combined seed as the basis for a pseudorandom number, and you could implement something like Civilization in a trustworthy manner without a central server of any sort.