..what library would make you interested eventually..

[Naughty Alien]

..just hypothetical question..lets say, if you are able to buy ready to use libraries, ready to use/plug in to LE, do you mind to score down in your posts, what would be that library from provided list, and eventually, if there is some other need of yours, with higher priority then one listed, just name it and shortly describe what it does...thanks guys

 

So here list go:

-Path finding library(including formations,hidding behind geometry, avoiding collisions)

-NPC Behaviour library

-Shader pack

-Camera control library(FPS, 3rd perspective, top down..pretty much everything exist with geometry avoiding and loads other stuff)

-Facial animation/expressions library

-LE-PS3 Translator library

-LE-Xbox360 Translator Library

-cut scene manager system(interactive/non interactive)

-..here fit anything of your needs..

 

Im just wondering what would be response (response can be also,'I dont need that junk , i can make it my own' ) .. so feel free and eventually add something you think its useful for ya thanks guys

[Rick]

I have to say a robust pathfinding system, hands down. Most every game in existence requires some form of pathfinding. After that would be shaders. A huge array of cool shaders can make a huge difference since they are so dang powerful.

[Pancakes]

1. Dialogue/Cutscenes Management (So you can direct and manage camera and decisions tree - think of Mass Effect2)

 

2. Pathfinding

 

3. Shaders

[gordonramp]

Lipsync animation,

Camera controls,

Pathfinding.

[ArBuZ]

Pathfinding. Absolutely agree with Rick. Almost every single game requires this. I would buy it if its not too expensive and is convenient to use in LE.

[Rick]

I would also pay for a pathfinding system in LE.

[goodone]

1. Camera control

2. Pathfinding

3. Shader pack

[Pixel Perfect]

Facial animation/expressions library would be of interest.

[franck22000]

PathFinding is a must

[Mumbles]

-LE-Xbox360 Translator Library

 

That's a big one for me... But would such a proposal require some sort of license (subscription to the XNA creators club perhaps)?

 

Lipsync animation

 

Definately agree with that one, but hopefully if this did get created, it would not need exceptional knowledge in 3ds to interact with it

 

 

 

(response can be also,'I dont need that junk , i can make it my own' )

 

I'll take you up on that

 

-Path finding library(including formations,hidding behind geometry, avoiding collisions)

I have to say a robust pathfinding system, hands down.

 

I think I'm going to be the first person to disagree with that one. Pathfinding in particular isn't necessarily one size fits all. Presumably this system would involve the use of A*? Personally, I would go with something that closely resembled Breadth-first search, but not identical because it does have a few minor hiccups. Hiding, and so on, I'm not against quite so much

 

For the record, I would say that NPC behaviour is even more to unique to each project, and so definitely something that done be done individually.

[Canardia]

- Pathfinding including Inverse Kinematics (so NPCs can climb over custom obstacles)

- Shaders

- 3D Particles including collision

- 3D Sounds including collision (and reflection)

[Rick]

I think I'm going to be the first person to disagree with that one. Pathfinding in particular isn't necessarily one size fits all. Presumably this system would involve the use of A*? Personally, I would go with something that closely resembled Breadth-first search, but not identical because it does have a few minor hiccups. Hiding, and so on, I'm not against quite so much

 

That's why I put Robust (always have an escape word ) I feel the library should include multiple styles on how it does it's implementation. I think the first criteria in deciding on answering his question "should" be what is most common in every game. From his list pathfinding is pretty high up there.

[omid3098]

-Cut scene manager system(interactive/non interactive)

-Shader pack

-Path finding library

-Facial animation/expressions library

-LE-PS3 Translator library

-LE-Xbox360 Translator Library

[paramecij]

- Animation & physics driven rag-dolls with IK and lip-sync support (definitely worth buying!!)

 

- Path-finding, yeah it's used in practically any game but it varies too much from game to game (might look at it if it's free, but probably wouldn't suit my needs for the price paid anyway)

[Rick]

- Animation & physics driven rag-dolls with IK and lip-sync support (definitely worth buying!!)

 

- Path-finding, yeah it's used in practically any game but it varies too much from game to game (might look at it if it's free, but probably wouldn't suit my needs for the price paid anyway)

 

I find other peoples view on this topic fascinating. To me the above doesn't make sense. I'm not saying you're wrong I'm just trying to figure out where you are coming from. I can't see lip-sync support or even rag-dolls being something that would take priority over a pathfinding system. Very few games actually use lip-sync technology. It's a cool feature but not exactly a requirement. Many games in the past have been very successful without it. On the other hand, even games made 20 years ago required some form of pathfinding. It's such a core game feature that it seems like the best candidate.

 

I also don't know if I buy the argument that there are so many kinds and nothing would fit every need. It seems to me that navmesh with A* has kind of taken over as the standard. Sure it might be overkill for some projects, but if it's already there why not use it? It provides pretty much all needs of any pathfinding system at pretty little performance cost.

 

I don't want to hijack the thread but I think this is kind of what the thread is about. Talking about these features and priorities. I'd be interested in hearing debates against pathfinding being a key component.

[Guest Red Ocktober]

yeah... pathfinding is almost a must have nowadays...

 

followed by an animation lib (with enhancements to enable facial expressions and talking)... enhanced physics (cloth and destructibles)... ehnhanced terrain (infinite streaming)...

 

i dunno what else... those are the first that come to mind...

 

 

--Mike

[Pixel Perfect]

I can't think of many games that don't require basic path finding as a minimum and do not see that as being game specific in any way. After all, if you want to move a NPC from a to b how else are you going to do it? Rick is right in as much that nav mesh node based systems utilising A* for the path finding are generally considered optimal these days ideally with full support for dynamic object avoidance and steering systems, but any form of path finding is preferable to none. Path finding can be as simple or as complex as you want to make it! I would place this as number one on the list.

 

Animation & physics driven rag-dolls with IK and lip-sync support .... great as a wish list but have you any idea how much work goes into this. Hope you have your credit card ready B)

 

I'd love to see some specifications for the systems you've developed so far Naughty Alien.

[Raul]

Assuming Pathfinding is just a piece of the AI I would say:

- AI: Pathfinding (environment collision detection + solve collision, friendly/enemy collision detection + solve collision, formation movement in RTS cases)

of course an RPG AI is differit from a RTS AI and also not the same with a FPS AI. B) But a Pathfinding system could be use in the same ways with the all types.

 

I would also pay for a good pathfinding lib but not thousands

[Mumbles]

I don't want to hijack the thread but I think this is kind of what the thread is about. Talking about these features and priorities. I'd be interested in hearing debates against pathfinding being a key component.

 

Then I'll start by ..trying.. to say why I don't think pathfinding is universal. (However, I'll never make it as a solicitor, as I'm not always brilliant at arguing my own case) Here goes:

 

Pathfinding is just that, finding a route from A to B. Obstacle avoidance, projectile avoidance, or similar do not count in my mind, as they would be better summed up using the words "obstacle avoidance" or "threat avoidance" and you get the idea. So once a route is found, the character sticks with it, unless its 'intelligence' tells it abort the path. Example being: An NPC in some sort of war game is walking towards a designated point holding a weapon as weak as a water pistol, and not in any sort of armoured transport. Suddenly it sees a sentry gun ahead. Knowing that it would be suicidal to continue in this direction, it looks around briefly, and decides to go somewhere else. Sadly the one size fits all pathfinding had no way of knowing that a static gun emplacement should be avoided in this situation - it was up to the soldier to see the sentry gun, and act on that, thus deciding to plot a new path to somewhere else.

 

How about a route being valid sometimes, and blatantly stupid at other times. Let's carry on with our same game that doesn't yet exist. Realising the battlefield is a dangerous place, our soldier has now hopped into a tank. Now our soldier has no fear about moving past machine guns since they are not going to inflict any damage. However, the route our algorithm found, tells our soldier to drive straight ahead .... into a river. Had the soldier have been on foot the route would have been valid, he could just swim across. Sadly tanks don't tend to float in water, so our poor soldier is wondering why his tank is stuck.

 

I'm sure given enough time I could think up more examples if neceaasry.

 

So whilst I'm agreeing that fathfinding is a key component, it has to integrate with the AI for it to work properly, and AI should never be an off the shelf product, therefore neither should pathfinding be.

[Rick]

I think you said it best when you said:

 

Pathfinding is just that, finding a route from A to B

 

That concept is universal and should be built into any engine. That's what I'm talking about. Then like you said, all the other stuff is AI related that is specific to the game. Pathfinding just gives you view of your available paths in data form. It would then be up to the AI specific code to use pathfinding to figure things out like you were describing.

 

My script attached to the soldier could be looking at every point in the path and seeing if there is a gunner in any sort of radius. If there is, it might try to find another path. All that stuff can be very specific but the general idea of finding a path from A to B is so generic that it should be part of any engine.

[Pixel Perfect]

Yes path finding is tied into AI but it's not the job of the path finder to act intelligently but rather to simply supply a viable path based on data supplied to it at the time the path is requested. It’s the AI which decides that a current path is no longer viable and requests a new path find at that point, possibly updating data the path finding is based on. It’s quite rare in a dynamic game for any lengthy path finds to be completed without being terminated well before this takes place due to changing circumstances.

 

The basic function of path finding is universal but needs to be based on dynamic data supplied by the AI system. Its ability to mimic real life successfully is of course a reflection of how well your AI performs.

[Mumbles]

But my point in the first example was, if the system had known not to take routes going past sentry guns, it would probably have found a better route, and not have to abandon the path due it being dangerous. if the the time the path is abandoned, and a new route is figured out to the same target, presumably the path is going to be the same. After all, is there a way of telling such a system "I don't want that route, I want another one to the same destination"? I would have though the re-searching would just search again and find the same path. Think of how much memory it is going to need to remember "bad" routes that it's already tried and found to be unsuccessful... May not be much on its own, but how many "good" and "bad" paths can be remembered at a time? As soon as a bad route is forgotten, it can be rediscovered.

[Pixel Perfect]

Yes, you can remove linked nodes or flag them unwalkable or use heuristics. This mimics how things are in real life. The first time you go into an unknown area you are totally relying on your senses to give you navigation feedback. You might know the rough direction you are going in or even have a distant visible target but your senses navigate you through the static geometry and alert you to dynamic threats such as the sentry gun.

 

The initial knowledge of the static geometry will enable the path finder to plot a course through that but it has no way of knowing about the sentry gun unless you've told it in advance (this might be true where the AI is simulating advanced reconnaissance). Therefore when the NPCs line of sight identifies and enemy sentry gun the AI would update the nodes surrounding that object and probably make them unwalkable, making all future path finds avoid that route. Heuristics can be used to make a route or area less likely to be chosen and may be set for example by the AI when enemy activity is detected in that area, it doesn't mean the path will never be chosen (you might have no choice if others are subsequently blocked) but whilst safer routes exist it would not be chosen.

[Rick]

Yeah, like Pixel is saying, you wouldn't store "bad" routes, you would update the grid. Either remove some nodes as walkable or make them very expensive to walk so the path finding system might not include it the next go around. I wouldn't want to remove it 100% because sometimes it might end up being the only way to get to point a (bridges out other routes or whatever) and so even though this would have a high weight to it, it would still be the only way.

 

The AI and pathfinding should be separate. They are standalone systems and should just have interfaces between each other.

[Pancakes]

someone made an open source pathfinder for blender game engine why don't you guys take a look at that

 

http://blenderartists.org/forum/showthread.php?t=158626

 

The Python script is right there for you to download and take a look at. Not sure if that would be helpful though. It's a bit over my head atm.