The actions the player can take in the game world are specified as very simple planning operators (see slide no. 11 of the game talkgame-talk.pdf). It is therefore a natural idea to really use the action schemata for planning, so that the game could plan sequences of actions that need to be taken to achieve a certain goal. There are two ways in which this might be useful.
The second subtask is to specify strategies for verbalizing the output of the planner. An important point that needs to be decided is how much of a plan should be told to the player. Some parts will have to be kept secret as it is part of the game that the player finds out certain things by himself. Here we might be able to start from work that has been done on verbalizing proofs (see e.g. Huang and Fiedler, 1997).
Preliminary. Will be updated shortly.
Xiaorong Huang and Armin Fiedler. Proof Verbalization as an Application of NLG. 15th International Joint Conference on Artificial Intelligence (IJCAI), 1997. Available at http://www.ags.uni-sb.de/~fiedler/home.html
Kristina Striegnitz kris@coli.uni-sb.de