Which one do you want to delete, if both are instanced dynamically ?
lol you must be precise.
Here is the scenario, object A has a collision detector. When a collision occurs, it gives you the name/reference of the other object, that we will name Object B.
Which one do you want to delete, A or B ?
the slower moving one?
if own['LastPos']!="Empty":
lastFrame=own['LastPos']
diff=lastFrame-own.worldPosition
Speed=diff.magnitude
own['LastPos']=own.worldPosition
Object b is the one i want to deleted.
Hi,
I am working on this, building a simple demo.
Had to add what hive is calling “bound object”.
I would call it an object graph.
In fact every graph is attached to an object now.
Now i create graph templates which are shared by the instances of the same GameObject.
This allows me to introduce a reference of any instanced game object into PlayNodes and set the proper collision events.
This is becoming really fun. Replicating the same kismet mechanisms.
What about specific nodes, like from the logic bricks?
Hey,
here it is, took me a while to add the inermediate nodes to achieve the collision itself.
Object A is the right cube, object b is each Bullet.
Collision event triggers deletion of bullets.
is python in between the nodes and the C++ level commands?
the logic bricks are C++ in the end right?
All is in python.
Exporting pynodes to xml, loading xml in the runtime, re-creating the graph structure and processing events with the python api.
I dont know how far it can get, but as long as things are possible in python, then it can be done with nodes.
And im not using sensor or actuators. lol.
Logic isnt what slows the bge the most, just to display these 2 cubes and a spot lamp, blender wants to die.
I would just delete the rasterizer and plug an ogre renderer behind, since many years there are the same bottlenecks and nobody seems to be working on optimization.
When you use custom material nodes, and press p you’ll play tomorow, today it will be compiling shaders or whatever. There is no hope. lol.
Its funnier to have a gameplay goal and create the required nodes to achieve it.
Do you want to see some specific gameplay ?
I want to try to make some animation blending, i dont know if the node editor is a good tool for that.
Are you going to replace the rasteriser with ogre renderer by yourself?
No im not crazy ! Just to understand what each files does it would take a year.
But it should be possible to create an addon that launches inside the viewport, not like ogrekit which remains external.
it would be worth it,
I don’t know why no one has done a better render yet,
I have a videogame idea that you could impliment with playnodes.You could make a enemy that punches the player with left and right fists in a programmed senquence.And the player would have to block the enemy with either shield on both his fists in order to take away the health of the enemy.So that he can kill the enemy.And the enemy will not hit the player when he has either shield up.You could make two enemies like this both with different programmed sequences of hitting with left or right
fists.
Haha street fighter 
I was thinking a openworld videogame like that.It would dynamically load terrain.But first you could if you like make the game
play.
Logic isnt what slows the bge the most, just to display these 2 cubes and a spot lamp, blender wants to die.
Yeah, im for Ogre3d too.
Its funnier to have a gameplay goal and create the required nodes to achieve it.
Nope that was just a general question. (i see a 5 nodes and from there comes the question)
Hi guys,
to have a simple fight system, you must create the proper logic to handle the various behaviors of the fighter.
Here is a drawing that explains the behavior of the fighter.
In the past i have done some simple finite state machine (FSM) fight system. Coding all by hand in python was laborious : https://www.youtube.com/watch?v=14RpTLVEg60
Now one need to build state machines with an editor, the base user shouldnt have to code anything.
Ive spent some time looking at how some engines perform animation blending.
In my opinion unity (mecanim) and ue4 (animation blueprint) are good ones for inspiration.
A character is a pupeet that you put in some temporary state untill some conditions happens, then you put him to another state, and this can go forever.
From what ive seen, a node editor is not a good solution to represent an FSM.
Because we want straith lines between nodes, we dont want the input/output socket mechanism, we dont even need the node properties. But we need to edit transitions from state to state, cliking on the lines that link nodes together. An we need to add conditions in each transition to switch between states.
I was looking at how to add a custom editor in blender, because this is a key element in a modern game engine.
So, i found this dcumentation wich explain in details how to create a custom editor in blender : http://wiki.blender.org/index.php/Dev:2.6/Source/Tutorials/AddAnEditor
Now this becomes serious.
For the sake of animation blending this dedicated editor is required.
We now have cycles baking since 2.71 and it performs at light speed, gives similiar or superior quality than lightmass or beast.
Depends on the number of samples, the integrator parameters (bounces), and the texture resolution of your lightmap.
So game developpment in blender becomes viable. I hope the rasterizer gets fixed soon too.
There is another way to do animation blending without a custm editor, using just a panel and common properties to setup the states and transitions.
And there are also some old methods using nodes like this. But you must setup everthing. An FSM editor is easier to use than this.
there is also this one http://udn.epicgames.com/Three/rsrc/Three/AnimationNodes/AnimNodeGameExample.jpg
my opinion, it is better to use nodes editor rather than adding FSM editor that takes more time to created.
There is a huge difference between an FSM editor and a (Behavior) Node editor.
You can compare that with a DSL and a general purpose language (like Python) or with a level editor and a game engine.
The FSM editor’s solely purpose is to describe an FSM as good as possible (focus is on FSM).
A node editor’s purpose is to describe behavior nodes (focus is on nodes).
They might look pretty similar, but they aren’t equal.
You can create an FSM with a behavior node editor (subset). But there is nothing that stops you from not creating an FSM.
With the FSM editor you always get an FSM (maybe not a good one, but an FSM). The FSM editor gives you all tools you need for an FSM. It does not distracts you with irrelevant options as a node editor does.
Example: You might notify the complains that BGE users see options that are relevant for Blender render but not for the BGE. It is the same dilemma. You want to create a game, but have to know what you can’t use, rather than you use everything you see.
BTW. I refer to the situation when you already want to describe an FSM.
I first thought about a json solution but when i looked at my screen, my hands already had coded an xml exporter, so xml imposed itself by speed
I love your style Your past with Kismet, programming skills and your confidence gives you good entry to improving the BGE features.
But please, don’t fixate on remaking the Kismet for BGE and look into the HIVE system. It’s been in development for long time and they could really use someone more people interested in developing node-based system. As far as I understand many points about it are still open and I’m sure you can help them shape it into something that you can all be proud of.
