OH MI GAWD
So interbutts, how are you today? I hope all is well in your chaotic realm.
So for my own personal notes:
A weather engine which uses pressure systems as moving nodes which gather in value to deter or manipulate cloud clusters which follow path seeking AI behavior.
To some of you, I bet you're going "what the fuck?" But here's how it works:
Pressure systems affect most movement of major weather and cloud clusters. Pressure systems are cold and hot air groups that move based on the rotation of the earth, geographic bodies, etc. That's simple enough and I'm sure you all knew that already. But, in what way could you represent this movement to simulate a real weather system on a simulated planet? You could do it with images sorta like how they do heightmaps to describe terrain. You could calculate it's movement. You could ignore it and throw weather systems all over the place. Or you could make a subset of your A.I. system, run pressure systems as entities which follow their own A.I. rules and the make another subset group representing cloud types that have avoidance and follow A.I. behavior based on the movement of pressure systems. In this way, you treat both pressure and clouds as entities which follow their own A.I. rules, and emulate real weather systems. Since A.I. algorithms allow for producedural generation, you could tweak the algorithm to be accurate enough to create their own effects. Volcanism, dust storms, snow storms, hurricanes, lightning, tornadoes, variable rain fall, heat lightning, etc. All of these effects could easily be represented following an A.I.-like system. We already know how to have A.I. activate behaviors. For example, a creature in a video game is aggro, and you get in range of it's aggression. It does a following A.I. until it's close enough and then runs an attack script. Easily, weather conditions could be represented, in an if/then statements or however. Either way, effects could be rendered through scripts and since scripts can be run real time, you could make changes on the fly and store it as behavior for specific planets based on their make up or what have you.
Another great thing about A.I. behavior is that it allows for procedural generation. There is such a thing as called "FLOCK" or "Boyd" behavior. Boyd is an alogorithm that allows you to have a series of entities follow each other in a group-like fashion. One unit goes one way, and the rest follow depending on the proximity to the unit, or other units of the group it is in. This is great because you can generate clouds on the fly, and use boyd behavior to display different clouds' movement. You can have variable cloud shapes and sizes that generate, cluster in a group, and then move along its generating A.I. path. You can even still use imposters to decrease the stress on the renderer, and also retrieve and store information about cloud movement in the heightmaps.
Sadly, I need more experience programming before I can take on a task like this, but it's a great idea I think.
another neat idea I have is about running a galaxy. Galaxies are huge, and there is no reason to run all the galactic information at once. If you're playing in a video game, cchances are there won't be hundreds of people on each planet. Even with WoW's numbers, you'd see maybe a few hundred on highly populate planets, but the majority of players will be out exploring on their own. However, it is known that if you were to run a galaxy chances are high that most players would seek out other players and thus stay in relatively high populated areas, and that there will be more places than there are people if rendered to scale. So, in that sense, you could easily make a heirarchical tree of parent/child node behavior that begins to organize players into areas. Players will be clustered in some areas and not in others. The areas they are not, you just shut down until someone is there. Since in a game as large as a galaxy, you'll have galactic events, and events all the way down to a microscopic level, you can initate events and then divide them as need be, in almost a quadtree or bintree or octtree like fashion. So, you'd have a galactic economy for instance. The economy at the galactic level can run on an algorithm and then work it's way down to a sector economy, then to a solar economy, and then finally to a planetary economy. All that would be needed is a rough estimate of occurrences, and then as child nodes are opened, you start to increase definition.
Did anyone see that study that came out explaining that solar systems like our own are rare? I think this is both disappointing and also wrong. I don't think we know enough yet about solar systems and galaxies to estimate the probability of occurrence. It did describe something we've commonly thought and know which is that planets have a tendency to form then spiral back inward into the star. I think one of the things we'll find is that young stars have a higher tendency to develop solid solar bodies like ours, and that older stars loose more of their original mass through formation and reformation. I think this because I believe both Jupiter and Saturn to be key planets in our formation. Even though the like to throw rocks at us, Jupiter's gravity throwing that one planet inbetween mars and jupiter out of sinc so it never formed could have been crucial in our stabalization. As well, it could be argued that jupiter throwing rocks may have slowed down or even stoped it's spiral inwards. And finally perhaps staurn kept big brother Jupiter out in the outer rim.
Does anyone think of Jupiter as the soviet russia planet of our solar system? I do. It's big, useless, assholish, and has tones of satelites. As well it's constantly throwing things into space without much thought. IN SOVIET JUPITER, SPACESHIP DRIVE YOU!
Finally, a website for myself:
http://rst.gsfc.nasa.gov/Sect14/Sect14_1c.html
Also cocks,
So for my own personal notes:
A weather engine which uses pressure systems as moving nodes which gather in value to deter or manipulate cloud clusters which follow path seeking AI behavior.
To some of you, I bet you're going "what the fuck?" But here's how it works:
Pressure systems affect most movement of major weather and cloud clusters. Pressure systems are cold and hot air groups that move based on the rotation of the earth, geographic bodies, etc. That's simple enough and I'm sure you all knew that already. But, in what way could you represent this movement to simulate a real weather system on a simulated planet? You could do it with images sorta like how they do heightmaps to describe terrain. You could calculate it's movement. You could ignore it and throw weather systems all over the place. Or you could make a subset of your A.I. system, run pressure systems as entities which follow their own A.I. rules and the make another subset group representing cloud types that have avoidance and follow A.I. behavior based on the movement of pressure systems. In this way, you treat both pressure and clouds as entities which follow their own A.I. rules, and emulate real weather systems. Since A.I. algorithms allow for producedural generation, you could tweak the algorithm to be accurate enough to create their own effects. Volcanism, dust storms, snow storms, hurricanes, lightning, tornadoes, variable rain fall, heat lightning, etc. All of these effects could easily be represented following an A.I.-like system. We already know how to have A.I. activate behaviors. For example, a creature in a video game is aggro, and you get in range of it's aggression. It does a following A.I. until it's close enough and then runs an attack script. Easily, weather conditions could be represented, in an if/then statements or however. Either way, effects could be rendered through scripts and since scripts can be run real time, you could make changes on the fly and store it as behavior for specific planets based on their make up or what have you.
Another great thing about A.I. behavior is that it allows for procedural generation. There is such a thing as called "FLOCK" or "Boyd" behavior. Boyd is an alogorithm that allows you to have a series of entities follow each other in a group-like fashion. One unit goes one way, and the rest follow depending on the proximity to the unit, or other units of the group it is in. This is great because you can generate clouds on the fly, and use boyd behavior to display different clouds' movement. You can have variable cloud shapes and sizes that generate, cluster in a group, and then move along its generating A.I. path. You can even still use imposters to decrease the stress on the renderer, and also retrieve and store information about cloud movement in the heightmaps.
Sadly, I need more experience programming before I can take on a task like this, but it's a great idea I think.
another neat idea I have is about running a galaxy. Galaxies are huge, and there is no reason to run all the galactic information at once. If you're playing in a video game, cchances are there won't be hundreds of people on each planet. Even with WoW's numbers, you'd see maybe a few hundred on highly populate planets, but the majority of players will be out exploring on their own. However, it is known that if you were to run a galaxy chances are high that most players would seek out other players and thus stay in relatively high populated areas, and that there will be more places than there are people if rendered to scale. So, in that sense, you could easily make a heirarchical tree of parent/child node behavior that begins to organize players into areas. Players will be clustered in some areas and not in others. The areas they are not, you just shut down until someone is there. Since in a game as large as a galaxy, you'll have galactic events, and events all the way down to a microscopic level, you can initate events and then divide them as need be, in almost a quadtree or bintree or octtree like fashion. So, you'd have a galactic economy for instance. The economy at the galactic level can run on an algorithm and then work it's way down to a sector economy, then to a solar economy, and then finally to a planetary economy. All that would be needed is a rough estimate of occurrences, and then as child nodes are opened, you start to increase definition.
Did anyone see that study that came out explaining that solar systems like our own are rare? I think this is both disappointing and also wrong. I don't think we know enough yet about solar systems and galaxies to estimate the probability of occurrence. It did describe something we've commonly thought and know which is that planets have a tendency to form then spiral back inward into the star. I think one of the things we'll find is that young stars have a higher tendency to develop solid solar bodies like ours, and that older stars loose more of their original mass through formation and reformation. I think this because I believe both Jupiter and Saturn to be key planets in our formation. Even though the like to throw rocks at us, Jupiter's gravity throwing that one planet inbetween mars and jupiter out of sinc so it never formed could have been crucial in our stabalization. As well, it could be argued that jupiter throwing rocks may have slowed down or even stoped it's spiral inwards. And finally perhaps staurn kept big brother Jupiter out in the outer rim.
Does anyone think of Jupiter as the soviet russia planet of our solar system? I do. It's big, useless, assholish, and has tones of satelites. As well it's constantly throwing things into space without much thought. IN SOVIET JUPITER, SPACESHIP DRIVE YOU!
Finally, a website for myself:
http://rst.gsfc.nasa.gov/Sect14/Sect14_1c.html
Also cocks,