Ceptre: A Language for Modeling Generative Interactive Systems.

Ceptre: A Language for Modeling Generative Interactive Systems.
Chris Martens

We present a rule specification language called Ceptre, intended to enable rapid prototyping for experimental game mechanics, especially in domains that depend on procedural generation and multi-agent simulation.

Ceptre can be viewed as an explication of a new methodology for understanding games based on linear logic, a formal logic concerned with resource usage. We present a correspondence between gameplay and proof search in linear logic, building on prior work on generating narratives. In Ceptre, we introduce the ability to add interactivity selectively into a generative model, enabling inspection of intermediate states for debugging and exploration as well as a means of play.

We claim that this methodology can support game designers and researchers in designing, anaylzing, and debugging the core systems of their work in generative, multi-agent gameplay. To support this claim, we provide two case studies implemented in Ceptre, one from interactive narrative and one from a strategy-like domain.

Some choice quotes from the artice follow.

Simple examples of the rule language:

The meaning of A -o B, to a first approximation, is that whenever the predicates in A are present, they may be replaced with B. One example of a rule is:

     at C L * at C’ L * likes C C’
  -o at C L * at C’ L * likes C C’ * likes C’ C.


Note that because of the replacement semantics of the rule, we need to reiterate everything on the right-hand side of the -o that we don’t want to disappear, such as the character locations and original likes fact. We use the syntactic sugar of prepending $ to anything intended not to be removed in order to reduce this redundancy:

do/compliment: $at C L * $at C’ L * $likes C C’ -o likes C’ C.

A more complex rule describes a murder action, using the ! operator to indicate a permanent state:

    anger C C’ * anger C C’ * anger C C’ * anger C C’
    * $at C L * at C’ L * $has C weapon
  -o !dead C’.

(This rule consumes C’s location, maintaining a global invariant that each character is mutually exclusively at a location or !dead.) Here we see a departure from planning formalisms: four instances of anger C C’ mean something different from one. Here we are using an emotion not just as a precondition but as a resource, where if we have enough of it, we can exchange it for a drastic consequence. Whether or not we diffuse the anger, or choose to keep it by prepending $ to the predicates, is an authorial choice.

Concurrency in narration:

Two rule applications that consume disjoint sets of resources from the same state can be said to happen concurrently, or independently. On the other hand, a rule that produces resources and another that consumes a subset of them can be said to be in a causal, or dependent, relationship. Less abstractly, if resources represent facts associated with particular game entities or characters, then independent rule applications represent potentially concurrent action by multiple such entities, and causally related rule applications represent either sequential action by a single actor, or synchronized interaction between two entities.

Stages, and a larger example:

We would like to for some of these rules to run automatically without player intervention. In our next iteration of the program, we will make use of a Ceptre feature called stages. Stages are a way of structuring a program in terms of independent components. Syntactically, a stage is a curly-brace-delimited set of rules with an associated name. Semantically, a stage is a unit of computation that runs to quiescence, i.e. no more rules are able to fire, at which point control may be transfered to another stage.


Additionally, we can test the design by “scripting” certain player strategies. For instance, we could augment the two rules in the fight stage to be deterministic, fighting when the monster can’t kill us in one turn and fleeing otherwise:

stage fight = {
      choice * $fight_in_progress * $monster Size * $health HP * Size < HP
    -o try_fight.
  do_flee :
      choice * fight_in_progress * $monster Size * $health HP * Size >= HP
    -o flee_screen.

If we remove interactivity from this stage, then we get automated combat sequences that should never result in the player’s death.