Adding Physics into Tremulous

[Nux]

there is almost no air control in vq3 physics

Almost being the operative word.

It's smaller, yes, but's it's still there and is definately noticable (such as when pouncing high with a dragoon). To remove that ability would make it harder to land on the exact right spot (like a moving human's head).

[Kaine]

Nux ninja-editted. 

[Urcscumug]

I'm not sure I understand exactly how this physics thing would improve Trem. I've noted some ideas, such as vehicles (elevators, cablecars) but I believe those are already possible; breakable stuff, which would be cool but not essential; stuff like having dead bodies retain shape and being able to block paths or climb over them, which is also cool but again not quite essential for playability.

So... can anybody give me some cold hard features that make physics absolutely must-have? I mean stuff that would tremendously improve the game of Trem instead of just the cool factor. (Which is not a small feat either, don't get me wrong, I'm all for more wow, just please make it clear what the goal is here.)

[Nux]

Nux ninja-editted. 

Prove it! xP

I'd also like to point out that if you're annoyed by dretches at your feet then imagine what adding super realistic physically interactive junk to your walkway would be like.

[Zero Ame]

I'm not sure I understand exactly how this physics thing would improve Trem. I've noted some ideas, such as vehicles (elevators, cablecars) but I believe those are already possible; breakable stuff, which would be cool but not essential; stuff like having dead bodies retain shape and being able to block paths or climb over them, which is also cool but again not quite essential for playability.

So... can anybody give me some cold hard features that make physics absolutely must-have? I mean stuff that would tremendously improve the game of Trem instead of just the cool factor. (Which is not a small feat either, don't get me wrong, I'm all for more wow, just please make it clear what the goal is here.)

The possiblities it would add is almost an endless list so I'm not going to go into THAT much detail, butit would be things like this:

You have a bunch of barrels and boxes all over the map, placers would be able to kick them around into a wall if they wanted to and funnel or block the enemy from getting to them temporarly until the enemy gets rid of the boxes and barrels(or destroys them if breakable stuff is added)

Which adds another point, you can't really have a breakable box unless there are physics to make it fall apart :/

Physics will not really improve or hinder gameplay really, its going to be more of a feature that will just be really awesome to have, thats basically all there is to it.

Maybe it will do something I'm not thinking of that would be really awesome, like maybe you could make rope ladders that actually sway back and forth when using them, or rope bridges that do the same thing, or hanging lights, ect, just think about it.

[gimhael]

In the Quake 3 engine all "objects" (like players, potenitally movable boxes etc.) are represented by AABBs (Axis Aligned Bounding Boxes), which makes it basically impossible to put a box in a 45 degree angle on the map. You can test this and build an armoury (doesn't look square) at an 45 degree angle, yet the box will still block players out of a perfectly square area aligned to the map walls.

Before you implement any kind of believable object motion/rotation, you will need a better collision model, and this in itself requires a lot of work (and will cost CPU cycles too).

[Amanieu]

Actually it should be relatively easy to add support for non-axis-aligned bounding boxes and allowing a non-square base. But this would require changes in the netcode and in the collision model on both client and server side.

[cactusfrog]

ya thats a good point the source engin is build off of q3 then we could code an engine a lot better.

[Amanieu]

You are clueless. The Source engine is closed source, so it won't help us. And it's based on q1, not q3.

[Kaleo]

You are clueless. The Source engine is closed source, so it won't help us. And it's based on q1, not q3.

I think you'll find that the Source engine is based of idTech2. Gold Source, the HL1 engine, is based of idTech1.

[Archangel]

source has nearly none of the original code anymore, anyway.

[==Troy==]

WolfET, jaymod.

[Kaleo]

UrT Has pretty good hitboxes.

[Plague Bringer]

Darkplaces engine, again (this time with Nexuiz): http://www.youtube.com/watch?v=kT4T0baxiu8&feature=related

[CATAHA]

We need at least breakable objects. =]

[Ivancool]

Tremulous Engine with New Physics = LAG!
Workaround=Limit servers up to 6 Players.
Result= :-(

[StevenM]

wolfET, also doesnt have lame unlagged.

[==Troy==]

wolfET, also doesnt have lame unlagged.

WolfET has unlagged by default. You just dont notice it.

[CATAHA]

Tremulous Engine with New Physics = LAG!
Workaround=Limit servers up to 6 Players.
Result= :-(

Most calculations can be prepared on local pc. Tremulous have many 'objects' in game right now and i heard not so many 'tremulous laggy!' opinions. So i really dont think that breakable objects can be SO laggy. What you only need - send ONCE to client small info that 'obj' broken. May be a bit more, but not so much. So i think you wrong with such opinion.

[Archangel]

wolfET, also doesnt have lame unlagged.

you don't notice it because it's real unlagged, not broken 1.1 unlagged.

[mooseberry]

wolfET, also doesnt have lame unlagged.

you don't notice it because it's real unlagged, not broken 1.1 unlagged.

Also the principles of unlagged work better when the opposing sides use the same types of weapons.

[Bissig]

wolfET, also doesnt have lame unlagged.

you don't notice it because it's real unlagged, not broken 1.1 unlagged.

As time goes by more and more in 1.1 seems horribly broken :-/

[Celestial_Rage]

wolfET, also doesnt have lame unlagged.

you don't notice it because it's real unlagged, not broken 1.1 unlagged.

As time goes by more and more in 1.1 seems horribly broken :-/

But we still love it. I will be hard to adjust to 1.2

[Winnie the Pooh]

Besides the obvious fun it would be to have physics in tremulous, lag would be a monumental problem and would probably take all the fun out of it. If a server is laggy enough, you will be able to walk right into the middle of an object, wait about half a second, and the object will then and only then react accordingly. This leads to a very boring or unsatisfactory performance of physics.

I am, however, all for breakables. Breakables look really nice (adding to the reputation of the game) ,they add to the strategy of the game, and also they perform according to each person's computer. The fragments of whatever object that is broken should not be user-interactive but rather they should "explode" in a random sequence. The parts should fly randomly.

[rotacak]

I think some people mixing "lag" and "low fps" together.

[BobTheDemonicTwit]

I think physics would creat too much  >:(LAG for Windows nubs

I use XP cuz I'm not allowed to change any "major" settings on my family's computer.

         Tryin to get laptop


a new OS isnt that "major" is it...

[mooseberry]

Getting a new OS will have pretty much zero affect on tremulous' performance. Don't sound so stupid, it hurts.

[Urcscumug]

But we still love it. I will be hard to adjust to 1.2

When I moved from 1.1 to TremFusion I didn't experience any bad stuff, only good things and none of them shocking. Is 1.2 supposed to be very different from Fusion?

[Amanieu]

You just tried the client part of tremfusion. The big difference in 1.2 is going to be game-side, such as the famous gameplay changes as well as some updated models, weapon animations, etc. You can try out the game part of Tremfusion on the Tremfusion test server.

[gimhael]

Besides the obvious fun it would be to have physics in tremulous, lag would be a monumental problem and would probably take all the fun out of it. If a server is laggy enough, you will be able to walk right into the middle of an object, wait about half a second, and the object will then and only then react accordingly. This leads to a very boring or unsatisfactory performance of physics.

The game engine already has to do collision detection anyway so that isn't a problem. And the collision detection is by far the largest part of the work, the code that causes the reaction to a collision is run only once when something is actually hit.

I am, however, all for breakables. Breakables look really nice (adding to the reputation of the game) ,they add to the strategy of the game, and also they perform according to each person's computer. The fragments of whatever object that is broken should not be user-interactive but rather they should "explode" in a random sequence. The parts should fly randomly.

Well, the game has some breakables, but they also have other functions (turrets, arms, etc.). When you add a few generic "buildables" (wall/box/stairs/whatever) which can only be placed by the map builder, then you have your breakables mod.