Request For Concept/model: Grappling Hook

[sparhawk]

Since the grappling hook comes practically free, in terms of coding, it would be nice if we could get a concept and a model for it. After all, why should we not include it when it can be already implemented with almost zero effort?

[Springheel]

Huh? Last I heard, there were still a number of hurdles to overcome before we have a fully working grappling hook. Are you saying all the advanced rope and collision physics have been worked out?

[oDDity]

It hardly requires a concept. A grappling hook is a grappling hook.

Ishtvan said he was going to play around with making a grappling hook, but I haven't heard anything abut it since.

edit: until I read Ishtvan's journal thread just now.

THe hard part of it will be making the af for it

[Macsen]

Would this be a grappling hook you shoot and sticks in walls like the rope arrow, or one you throw and grabs onto ledges?

[sparhawk]

Obviously a grappling hook wouldn't stick into walls. And Yes. Ishtvan says that he apparently fixed the major problems with the rope.

[Macsen]

Obviously a grappling hook wouldn't stick into walls. And Yes. Ishtvan says that he apparently fixed the major problems with the rope.

 

There are two types of grappling hook. The first is the standard burglar type, like this.

 

The second type shoots straight forward and slams into the wall. Unlike the rope arrow the player remains attached and is tugged after the rope/chain, which retracts into the device.

[sparhawk]

We only want a grappling hook that the player can use to climb. No automatich dragging.

[oDDity]

Cool as that burglary sequence at the start of The Pink Panther was, it's a bit too modern for this game.

[sparhawk]

What we might consider though, is that the rope of the grappling hook can be attached to some geometry, so that the player can climb stright over some gap. For example. The player could climb a tower and then throws the hook on the next tower and attaches the rope to the place where he stands.

[Ishtvan]

It doesn't need to be an AF for now, one can just shoot a rope arrow into it to attach a rope to it for testing(although dragging it back toward you currently won't work to move the grapple in that case, needs some coding). It needs to have a pretty good collision model though if we want it to actually work by catching on things.

[oDDity]

Couldn't you make it use the model's polygons as the collision model. THne you wonuldn't even need any kind of fancy skeleton or AF, just use the pronged head of the model.

[Ishtvan]

For idMovables I think you actually need a texture with common/clip on it to act as the clipmodel. Otherwise it auto-generates one and the results may be crap. If you create a slightly larger shell following the shape of the model and textured it with common/clip though, that would be essentially using the polies of the model for collion.

[Domarius]

Well ideally the CM "shell" should have less polys - That's the point of the CMs. More faces means more CPU processing time, which is mostly unnessecary. You just need the basic form.

[Ishtvan]

For a grappling hook to actually work as intended though we need more than a cube around it. We need to model the inner faces so it can hook onto things.

[sparhawk]

A cube would be no point, but a CM that follows the basic outline of the hooks without the details should be fine.

[oDDity]

I think there's a limit on how many faces can be in a CM anyway.

[Ishtvan]

I doubt the limit would be hit when making the hooks though; some of the .CM's, like a "round" barrel have quite a lot of faces.

 

You're right that we'll eventually want to do it as an AF though, unless we can do something like bind an idChain AF to the grappling hook, then we'd have the grapple plus a dynamic amount of rope. Entity binding has some problems though.

[Domarius]

lol I didn't mean a cube! May as well do AABB if I meant a cube

 

A SIMPLIFIED version of the original model.

 

For example you don't need to test for collision with peoples individual nostrils, and probably not even their nose - you only want to see if a collision happened with the general surface of their face.

 

Same goes for individual fingers - they are so close together and take up so many polys that when we're talking about poygon collision, it just is so not worth the CPU time.

[Ishtvan]

I'm just saying a grappling hook needs to have at least the hooks modeled in the CM otherwise it won't work right. Of course we don't need every nook and cranny.