ut2004 on godot question?

[Anonymous.]

The vision would be to re-implement our game with something we can continue to support in the future (not something like UE5). If Omni, HoC, SoL, etc are alive, it is because of all the mods. It would be nice to be able to actually fix bugs and support the modding effort going forward (less gray hair for Snarf too). This is why I suggested Godot. No more getting locked into these jackass companies.

Does godot use a similar directory and dependency system like UT2004? UT4 allowed assets to be baked into the map, and baked maps were impossible to edit (intentional). I don't even know if it was possible to link to global dependencies, but if it was, they set the precedent to the opposite of that. There was no scripting language besides C++, as with blueprint you couldn't edit anything that wasn't already a variable. All of these intentional features were excellent in preventing modding, as at the time epic was already against letting games live past their useful lives.

[pooty]

Funny that you bring up ML. I would be fun to train a few bots following e.g., Enyo styles, mid-level, and low-skilled players. Then when Enyo joins in a middle of a match, you could balance it with an "Enyo" bot. A harder goal would be making an effective bot that left the fun vehicles for the human players. We'd have to make Enyo only use scorpions, raptors, and goliaths for training purposes. We could do all this stuff if we had more control.

UT2004 bots already have varying skill levels. Xan anyone? You can spawn Actors in the middle of match (see Alien2)...not sure it'd work with bots as they are considered players. There's some basic bot pathing and vehicle use in there already.

[captainsnarf]

I came across an arenaFPS that was developed by a couple guys about a year back. It took them two years to get a basic DM type game working. Two years! But it looked very good, similar to UT4 (alpha) game. I asked them how they did all programming, modelling, etc. They said they didn't. They did most of the programming, but bought all of the art and 3D models from contractors. They spent about $80k of their own savings buying 3D models and textures.

Unless we want to try and sell our game, I don't see any problem just continuing with modding UT2004 forever. There is already a mutator for the server list issue. People are working on new patches also. The game is removed from stores, but will always be available on e.g. torrent. I'm sad there won't be any new players, but we really need a new UT game for that.

For a completely free game with source code, GPL ,etc, I definitely recommend xonotic as the way forward. A part of UTComp is the 'forward' mutator which turns UT into more of a Quake game. I see no reason why we couldn't do the same to a quake clone like xonotic - add a 'tournament' mod to xonotic that turns it into UT2004. The main issue is lack of vehicles, but I'd be happy with dodge jump and shock rifles tbh.

[YEAAAHHHHHHHHHH]

Funny that you bring up ML. I would be fun to train a few bots following e.g., Enyo styles, mid-level, and low-skilled players. Then when Enyo joins in a middle of a match, you could balance it with an "Enyo" bot. A harder goal would be making an effective bot that left the fun vehicles for the human players. We'd have to make Enyo only use scorpions, raptors, and goliaths for training purposes. We could do all this stuff if we had more control.

UT2004 bots already have varying skill levels. Xan anyone? You can spawn Actors in the middle of match (see Alien2)...not sure it'd work with bots as they are considered players. There's some basic bot pathing and vehicle use in there already.

The "BadAzz GitRDone" Ballistic server might do something like this. Or maybe it's just by name. But their bots are named after regular players and then have a server tag at the end of the name.

[FuriousRabbit]

The vision would be to re-implement our game with something we can continue to support in the future (not something like UE5). If Omni, HoC, SoL, etc are alive, it is because of all the mods. It would be nice to be able to actually fix bugs and support the modding effort going forward (less gray hair for Snarf too). This is why I suggested Godot. No more getting locked into these jackass companies.

Does godot use a similar directory and dependency system like UT2004? UT4 allowed assets to be baked into the map, and baked maps were impossible to edit (intentional). I don't even know if it was possible to link to global dependencies, but if it was, they set the precedent to the opposite of that. There was no scripting language besides C++, as with blueprint you couldn't edit anything that wasn't already a variable. All of these intentional features were excellent in preventing modding, as at the time epic was already against letting games live past their useful lives.

I'm not familiar enough the ut2004 code to fully answer. But compared to UT4, it wouldn't matter because all the source needed to run the game would be available. The Godot editor would probably be the editor for building a new map or vehicle. Still reading a lot but... a Godot game is composed of "scenes". Each scene is composed of a number of nodes. So a tank might be a scene, and have nodes representing the geometry, hit box, etc. The root node on each scene is usually accompanied with scripting where you bring in all your signals, write custom functions, or extend built-in ones.

You can write the script in either GDScript, C#, or C++ (somehow). I'd imagine some of the higher level code might be C++ (netcode, threading, etc), but most of the objects in the game would be GDScript. It's quite accessible. Adding a new vehicles or maps would be straightforward - copy a scene and customize. But you could also create new capabilities in an unrestricted way, even if it mean modifying the core scripts and structures.

[FuriousRabbit]

Funny that you bring up ML. I would be fun to train a few bots following e.g., Enyo styles, mid-level, and low-skilled players. Then when Enyo joins in a middle of a match, you could balance it with an "Enyo" bot. A harder goal would be making an effective bot that left the fun vehicles for the human players. We'd have to make Enyo only use scorpions, raptors, and goliaths for training purposes. We could do all this stuff if we had more control.

UT2004 bots already have varying skill levels. Xan anyone? You can spawn Actors in the middle of match (see Alien2)...not sure it'd work with bots as they are considered players. There's some basic bot pathing and vehicle use in there already.

Yes but usually increasing skill means making them super human in some ways while remaining super stupid in other ways.

[FuriousRabbit]

I came across an arenaFPS that was developed by a couple guys about a year back. It took them two years to get a basic DM type game working. Two years! But it looked very good, similar to UT4 (alpha) game. I asked them how they did all programming, modelling, etc. They said they didn't. They did most of the programming, but bought all of the art and 3D models from contractors. They spent about $80k of their own savings buying 3D models and textures.

My hope is matching nearly 20-year-old graphics tech would be cheap or free. The focus would be on gameplay since this crowd is all happy with a 20-year-old game anyways. If it was successful, I'm sure someone would make it fancy later (see trans_lux's post of what looks like ray tracing in ut). Right now I'm thinking 1-year to freon, which probably means 2 years

Unless we want to try and sell our game, I don't see any problem just continuing with modding UT2004 forever. There is already a mutator for the server list issue. People are working on new patches also. The game is removed from stores, but will always be available on e.g. torrent. I'm sad there won't be any new players, but we really need a new UT game for that.

The crashes and performance issues (in a 20-year-old game!) are ongoing problems. Selling it is a high bar. Better to assume it wouldn't fair better than toxxik.

For a completely free game with source code, GPL ,etc, I definitely recommend xonotic as the way forward. A part of UTComp is the 'forward' mutator which turns UT into more of a Quake game. I see no reason why we couldn't do the same to a quake clone like xonotic - add a 'tournament' mod to xonotic that turns it into UT2004. The main issue is lack of vehicles, but I'd be happy with dodge jump and shock rifles tbh.

Godot has (my estimate) 2-orders of magnitude more contributors. Much safer bet. They also support everything including Android, iOS, etc. The next version is heavily focused on Vulkan. Probably no chance of much of that with xonotic. I'm actually quite impressed with Godot. Had no issues whatsoever building it from source and found the documentation excellent. Also, at least for me, Godot is a resource I can apply to other goals (not necessarily games). That's probably true for a lot of other people too. I would also use GPL as a license so we could pull resources from xonotic, and they could pull from us.

[captainsnarf]

I came across this yesterday

https://www.rigsofrods.org/

Rigs of Rods is a free and open source vehicle simulator that uses soft-body physics to simulate the motion destruction and deformation of vehicles

It would be cool to integrate this with a godot fps. It's currently built around OGRE but shouldn't be too hard to port. I like that it already has scripting, an editor, and multiplayer.

[FuriousRabbit]

Yeah, that would be cool. The scorpions always were a little squirrely. Looks like OGRE is just a renderer so all the physics would be in rigsofrods.

I managed to pull the character meshes and animations from ut2004. Working on a pathfinder to see what is possible.

[pooty]

I managed to pull the character meshes and animations from ut2004. Working on a pathfinder to see what is possible.

Nice

Seems like we have a few people willing to explore. It would be really cool, but a bit of dream, to get a group together and make it actually work. I think we have decent enough coding skills here to do it.