world

World Sub-Project

Objective

The world sub-project shall create a framework which enables programmers to quickly and esily build virtual worlds of very different kinds as playground for very different agents. This includes possiblitities for agents to interact with the world, simple physical models and dynamic procedures within the world.

Part of the work is to implement one or more sample worlds - also a world which will be designed by the esthetics sub-project - and to support other people who want to create their own worlds.

This sub-project also contributes a world viewer/editor to the toolkit sub-project, which can be used to watch, change or create a simulated world.

Status

• Static world works
• Agent interaction works
• There is some support for dynamic actions, but largly untestet and sometimes unfinished. This has to be extended and cleaned up
• World physics needs to be written/rewritten
• World editor works, but there are a *lot* of things of things which would be nice to have ;-)

Current Activity & Discussion

• Doerner World Concept
• big todo: Create some nicely designed, consistent worlds. For discussion about about this, see Possible Worlds. For the current concept for a Doerner-like world, see new world.
• hierarchical objects: While we need "high level" objects (let's say "a tree") for world simulation, editor/viewer and the 3dviewer sub-project, agents should be able to make their own assumptions concerning the grouping of their sensations/possible targets for actions. Therefore, it will be made possible to build hierarchies of objects (a tree consists of a trunk and a crown, which consists of leaves), and to show to agents only the leaves of the tree (figuratively speaking ;-) ) as possible action targets and handle their actions accordingly, while the rest of the framework works with the tree as a whole thing. Implementation of this is nearly done.
• make it possible to switch world dynamics on or off / different levels of dynamics. Todo: What dynamic events should occour? When?
• partly done: make it possible to show current agent actions / mental activities? in world editor

Feature History

WorldEditor

0.6.7

• select objects using selection rectangle
• new method for creating objects

• can select multiple objects
• zooming scales also icons, not only space between them